refactor(subagent): shorten model-selection tool descriptions

docs/tool-catalog.i18n.yaml

@@ -225,11 +225,11 @@
   en: 841abe0dd5913ea7
   zh: 841abe0dd5913ea7
 /tool-schema-catalog/deepseek-ai-dsh-tool-subagent/list-subagent-models:
-  en: aba834ab4f70005a
-  zh: 142f819e8da5acc7
+  en: 6872bee11fccb6c8
+  zh: 230f2d443dc5d9a6
 /tool-schema-catalog/deepseek-ai-dsh-tool-subagent/subagent:
-  en: 0de8d49aa443ec9c
-  zh: 7cb7c8a55a0469b1
+  en: aff4a6668b41b7d5
+  zh: 98d6ddf8aa28c7f9
 /tool-schema-catalog/deepseek-ai-dsh-tool-subagent-control:
   en: ab82e130c3413f04
   zh: ab82e130c3413f04

docs/tool-catalog.md

@@ -2017,7 +2017,7 @@ The five read-only tools hide provider cursors and authorize every result from t
 
 ### `list_subagent_models`
 
-Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.
+Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.
 
 ```json
 {
@@ -2025,11 +2025,11 @@ Discover LLM routes for subagents without changing the current Agent. Call with
   "properties": {
     "provider": {
       "type": "string",
-      "description": "Registered LLM provider id. Omit to list providers."
+      "description": "Registered LLM provider id."
     },
     "model": {
       "type": "string",
-      "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+      "description": "Exact model id; requires provider."
     }
   }
 }
@@ -2039,7 +2039,7 @@ Source: [`packages/subagent/tool-subagent/src/list-models.ts`](../packages/subag
 
 ### `subagent`
 
-Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.
+Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.
 
 ```json
 {
@@ -2059,15 +2059,15 @@ Delegate a self-contained task to a subagent (a separate agent that works in its
     },
     "provider": {
       "type": "string",
-      "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+      "description": "Child LLM provider id."
     },
     "model": {
       "type": "string",
-      "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+      "description": "Model id within that provider."
     },
     "reasoning_effort": {
       "type": "string",
-      "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+      "description": "Model-specific reasoning effort id."
     }
   },
   "required": [

docs/tool-catalog.zh.md

@@ -2023,7 +2023,7 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后，
 
 ### `list_subagent_models`
 
-发现 subagent 可用的 LLM 路由，不更改当前 Agent。无参数调用会列出已注册提供方；提供 `provider` 时会列出其允许的模型；同时提供 `provider` 和 `model` 时会检查该精确模型及其推理强度。目录条目只提供建议：adapter 可能接受未列出的模型 id。把返回的 id 用于委派工具的 `provider`、`model` 与 `reasoning_effort` 字段。
+查询 subagent 可用的模型，不更改当前 Agent。无参数：提供方；`provider`：允许的模型；`provider` + `model`：该模型及其推理强度。将返回的 id 用于委派。
 
 ```json
 {
@@ -2031,11 +2031,11 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后，
   "properties": {
     "provider": {
       "type": "string",
-      "description": "Registered LLM provider id. Omit to list providers."
+      "description": "Registered LLM provider id."
     },
     "model": {
       "type": "string",
-      "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+      "description": "Exact model id; requires provider."
     }
   }
 }
@@ -2045,7 +2045,7 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后，
 
 ### `subagent`
 
-将一项自包含任务委派给 subagent（在自身上下文中工作的独立 agent），用它卸载聚焦且独立的工作，例如研究、限定范围的实现或分析，以免消耗当前对话的上下文。subagent 会返回结果，但不会返回中间步骤。此工具启动独立管理的 subagent 并立即返回其 id。运行时会在任务结束时通知你。子代理通过 `send_message` 汇报结果；你可以使用 `send_message` 在运行时引导它，或在结束后继续对话。子级 LLM 选择是可选的。省略 `provider`、`model` 和 `reasoning_effort` 会使用配置的子级默认值，并从父 Agent 继承缺失且兼容的值。先通过 `list_subagent_models` 查询路由和强度，再同时提供 `provider` 与 `model`。更改生效路由却不指定强度时，使用所选模型的默认强度。
+将一项自包含任务委派给 subagent（在自身上下文中工作的独立 agent），用它卸载聚焦且独立的工作，例如研究、限定范围的实现或分析，以免消耗当前对话的上下文。subagent 会返回结果，但不会返回中间步骤。此工具启动独立管理的 subagent 并立即返回其 id。运行时会在任务结束时通知你。子代理通过 `send_message` 汇报结果；你可以使用 `send_message` 在运行时引导它，或在结束后继续对话。先通过 `list_subagent_models` 查询选项，再同时设置 `provider` 与 `model`。省略的值使用子级默认值，缺失时使用兼容的父级值。生效路由改变时，省略 `reasoning_effort` 会使用模型的默认强度。
 
 ```json
 {
@@ -2065,15 +2065,15 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后，
     },
     "provider": {
       "type": "string",
-      "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+      "description": "Child LLM provider id."
     },
     "model": {
       "type": "string",
-      "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+      "description": "Model id within that provider."
     },
     "reasoning_effort": {
       "type": "string",
-      "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+      "description": "Model-specific reasoning effort id."
     }
   },
   "required": [

packages/subagent/tool-subagent/src/index.ts

@@ -210,8 +210,8 @@ export function apply(ctx: Context, config: Config, session?: Session): void {
       const wording = providerWording(subagentProvider.inheritsParentContext)
       const providerRouteDefaults = subagentProvider.agentRouteDefaults
       const selectionDescription = providerRouteDefaults !== undefined
-        ? ' Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and this provider\'s route defaults. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model\'s default effort.'
-        : ' Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model\'s default effort.'
+        ? ' Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to this backend\'s route defaults. When the effective route changes, an omitted `reasoning_effort` uses the model\'s default.'
+        : ' Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model\'s default.'
       const definition = defineTool({
         name: toolName,
         description: wording.description
@@ -238,21 +238,15 @@ export function apply(ctx: Context, config: Config, session?: Session): void {
           ...modelSelectionEnabled ? {
             provider: {
               type: 'string' as const,
-              description: providerRouteDefaults !== undefined
-                ? 'LLM provider route for the child. Supply together with model; omit both to use configured child defaults or this provider\'s route defaults.'
-                : 'LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route.',
+              description: 'Child LLM provider id.',
             },
             model: {
               type: 'string' as const,
-              description: providerRouteDefaults !== undefined
-                ? 'Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or this provider\'s route defaults.'
-                : 'Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route.',
+              description: 'Model id within that provider.',
             },
             reasoning_effort: {
               type: 'string' as const,
-              description: providerRouteDefaults !== undefined
-                ? 'Adapter-owned reasoning effort for the effective child route. Omit to use a compatible configured effort or the selected model\'s default.'
-                : 'Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model\'s default.',
+              description: 'Model-specific reasoning effort id.',
             },
           } : {},
 

packages/subagent/tool-subagent/src/list-models.ts

@@ -104,19 +104,17 @@ export function registerListSubagentModels(ctx: Context, policyFor: (exec: ToolR
   ctx.tools.register(defineTool({
     name: 'list_subagent_models',
     description:
-      'Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list '
-      + 'registered providers, with `provider` to list its allowed models, or with `provider` and `model` '
-      + 'to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may '
-      + 'accept an unlisted model id. Use the returned ids with a delegation tool\'s `provider`, `model`, and '
-      + '`reasoning_effort` fields.',
+      'Inspect subagent models without changing the current Agent. No arguments: providers; '
+      + '`provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. '
+      + 'Use returned ids for delegation.',
     parameters: {
       provider: {
         type: 'string',
-        description: 'Registered LLM provider id. Omit to list providers.',
+        description: 'Registered LLM provider id.',
       },
       model: {
         type: 'string',
-        description: 'Exact model id to inspect. Requires provider; omit to list that provider\'s allowed models.',
+        description: 'Exact model id; requires provider.',
       },
     },
     output: {

packages/subagent/tool-subagent/tests/tool-subagent.spec.ts

@@ -266,7 +266,7 @@ describe('dsh-tool-subagent', () => {
       model: 'child-model',
     })
     expect(ctx.tools.schemas(modelSelectionSetupAgent(ctx)).find(schema => schema.name === 'subagent')?.description)
-      .toContain('this provider\'s route defaults')
+      .toContain('this backend\'s route defaults')
     expect(seen?.agentOptions).toEqual({
       provider: 'alpha',
       model: 'child-model',
@@ -513,7 +513,7 @@ describe('dsh-tool-subagent', () => {
     const schema = ctx.tools.schemas().find(s => s.name === 'subagent')!
     expect(schema.description).toContain('inherits this conversation')
     expect(schema.description).not.toContain('does not see this conversation')
-    expect(schema.description).not.toContain('Child LLM selection')
+    expect(schema.description).not.toContain('Inspect options with `list_subagent_models`')
     const props = (schema.parameters as { properties: Record<string, { description: string }> }).properties
     expect(props['prompt']!.description).toContain('completed turns')
   })

snapshots/sdk/bash-tool/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-activation-limit/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-continuable-inheritance/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-continuable/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-dsh-sdk-diagnostic/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -393,7 +393,7 @@
     },
     {
       "name": "subagent_dsh_sdk",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and this provider's route defaults. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to this backend's route defaults. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -411,15 +411,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to use a compatible configured effort or the selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -430,7 +430,7 @@
     },
     {
       "name": "subagent_dsh_sdk_second",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and this provider's route defaults. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to this backend's route defaults. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -448,15 +448,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to use a compatible configured effort or the selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-dsh-sdk-dynamic-route/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-dsh-sdk-dynamic-route/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and this provider's route defaults. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to this backend's route defaults. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or this provider's route defaults."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to use a compatible configured effort or the selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-list-agents/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/subagent-send-message/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/text-turn/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/sdk/tool-error-details/system-prompt.expected.md

@@ -173,11 +173,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -206,7 +206,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -214,11 +214,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/sdk/tool-error-details/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/advanced-toolchain/system-prompt.expected.md

@@ -186,11 +186,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -219,7 +219,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -227,11 +227,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/advanced-toolchain/tool-schemas.expected.json

@@ -303,17 +303,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -397,7 +397,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -415,15 +415,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -1045,17 +1045,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1139,7 +1139,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1157,15 +1157,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [

snapshots/session/agent-instructions/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -924,17 +924,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1018,7 +1018,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1036,15 +1036,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [

snapshots/session/background-confinement-failure/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/both-mode-turn/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/browser-use-chrome-devtools-mcp/tool-schemas.expected.json

@@ -302,17 +302,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -463,7 +463,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -481,15 +481,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/browser-use-playwright-mcp/tool-schemas.expected.json

@@ -302,17 +302,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -465,7 +465,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -483,15 +483,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/browser-use-stagehand-native/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -569,7 +569,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -587,15 +587,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/claude-code-mods/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/compaction-recovery/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -924,17 +924,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1018,7 +1018,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1036,15 +1036,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [

snapshots/session/computer-use-cua-driver-mcp/tool-schemas.expected.json

@@ -302,17 +302,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -442,7 +442,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -460,15 +460,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/computer-use-cua-driver-native/tool-schemas.expected.json

@@ -308,17 +308,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -402,7 +402,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -420,15 +420,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/cordis-inspect-jsdoc/tool-schemas.expected.json

@@ -303,17 +303,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -397,7 +397,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -415,15 +415,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/deepseek-messages-system-prompt/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/deepseek-protocol-system-prompt/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/dynamic-tool-updates/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -924,17 +924,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1026,7 +1026,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1044,15 +1044,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [
@@ -1593,17 +1593,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1687,7 +1687,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1705,15 +1705,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [

snapshots/session/foreground-confinement-timeout/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/fs-glob-sampling/tool-schemas.expected.json

@@ -182,17 +182,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -260,7 +260,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -278,15 +278,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/lsp-definition/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -393,7 +393,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -411,15 +411,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/mcp-resources-ptc/system-prompt.expected.md

@@ -162,11 +162,11 @@ interface ToolArgsMap {
     /** Continuation cursor returned by this server. */
     cursor?: string;
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -202,7 +202,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -210,11 +210,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/mcp-resources/tool-schemas.expected.json

@@ -302,17 +302,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -417,7 +417,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -435,15 +435,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/multimodal-spill-ends/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -375,7 +375,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -393,15 +393,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/office-progressive-native/tool-schemas.expected.json

@@ -217,17 +217,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -311,7 +311,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -329,15 +329,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/office-progressive-ptc/system-prompt.expected.md

@@ -119,11 +119,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -152,7 +152,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -160,11 +160,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/office-skills/tool-schemas.expected.json

@@ -217,17 +217,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -311,7 +311,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -329,15 +329,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/persistent-pwsh-tool-turn/tool-schemas.expected.json

@@ -137,17 +137,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -231,7 +231,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -249,15 +249,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/plugin-manager-mcp/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -417,7 +417,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -435,15 +435,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [
@@ -1025,17 +1025,17 @@
       },
       {
         "name": "list_subagent_models",
-        "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+        "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
         "parameters": {
           "type": "object",
           "properties": {
             "provider": {
               "type": "string",
-              "description": "Registered LLM provider id. Omit to list providers."
+              "description": "Registered LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+              "description": "Exact model id; requires provider."
             }
           }
         }
@@ -1210,7 +1210,7 @@
       },
       {
         "name": "subagent",
-        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+        "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
         "parameters": {
           "type": "object",
           "properties": {
@@ -1228,15 +1228,15 @@
             },
             "provider": {
               "type": "string",
-              "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+              "description": "Child LLM provider id."
             },
             "model": {
               "type": "string",
-              "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+              "description": "Model id within that provider."
             },
             "reasoning_effort": {
               "type": "string",
-              "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+              "description": "Model-specific reasoning effort id."
             }
           },
           "required": [

snapshots/session/plugin-manager/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -417,7 +417,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -435,15 +435,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/product-subagent-both/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/product-subagent-codex/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/product-subagent-result-diagnostic/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/provider-cwd/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/ptc-node-read-only/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/ptc-node-workspace/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/ptc-persistent/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/ptc-policy-change/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/ptc-python-turn/system-prompt.expected.md

@@ -288,9 +288,9 @@ class ListAgentsOutput2(TypedDict):
     depth: NotRequired[float]
 
 class ListSubagentModelsArgs(TypedDict):
-    # Registered LLM provider id. Omit to list providers.
+    # Registered LLM provider id.
     provider: NotRequired[str]
-    # Exact model id to inspect. Requires provider; omit to list that provider's allowed models.
+    # Exact model id; requires provider.
     model: NotRequired[str]
     # Additional keys beyond those declared are allowed.
 
@@ -377,11 +377,11 @@ class SubagentArgs(TypedDict):
     description: str
     # The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs.
     prompt: str
-    # LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route.
+    # Child LLM provider id.
     provider: NotRequired[str]
-    # Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route.
+    # Model id within that provider.
     model: NotRequired[str]
-    # Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default.
+    # Model-specific reasoning effort id.
     reasoning_effort: NotRequired[str]
     # Additional keys beyond those declared are allowed.
 
@@ -549,7 +549,7 @@ class Tools(Protocol):
     async def list_agents(self, args: ListAgentsArgs) -> list[ListAgentsOutput1 | ListAgentsOutput2]:
         """List subagents you started, with their ids, labels, and status. running means it is working; inactive means it is not currently working. You will be notified when a subagent finishes; there is no need to keep checking its status. Use send_message to continue the conversation."""
     async def list_subagent_models(self, args: ListSubagentModelsArgs) -> str:
-        """Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields."""
+        """Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation."""
     async def read(self, args: ReadArgs) -> ReadOutput:
         """Read a UTF-8 text file and return line-numbered content."""
     async def read_image(self, args: ReadImageArgs) -> ReadImageOutput:
@@ -559,7 +559,7 @@ class Tools(Protocol):
     async def skill(self, args: SkillArgs) -> SkillOutput:
         """Load the full instructions for a skill. Call it before acting on a task that names or clearly matches a skill in the session skill catalog."""
     async def subagent(self, args: SubagentArgs) -> SubagentOutput:
-        """Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort."""
+        """Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default."""
     async def subagent_fork(self, args: SubagentForkArgs) -> SubagentForkOutput:
         """Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes."""
     async def todo_write(self, args: TodoWriteArgs) -> TodoWriteOutput:

snapshots/session/ptc-read-image/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/ptc-turn/system-prompt.expected.md

@@ -140,11 +140,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Read a UTF-8 text file and return line-numbered content. */
@@ -173,7 +173,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -181,11 +181,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/session/pty-tools-sandbox-backend/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/pwsh-tool-turn/tool-schemas.expected.json

@@ -189,17 +189,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -312,7 +312,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -330,15 +330,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/ralph-loop/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -419,7 +419,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -437,15 +437,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/ralph-loop/tool-schemas.2.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -419,7 +419,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -437,15 +437,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/ralph-loop/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -376,7 +376,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -394,15 +394,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/session-query-spill/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -560,7 +560,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -578,15 +578,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-acp-diagnostic/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-child-question-rejection/tool-schemas.expected.json

@@ -325,17 +325,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -419,7 +419,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -437,15 +437,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-no-message-tool/tool-schemas.1.expected.json

@@ -176,17 +176,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -249,7 +249,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -267,15 +267,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-structured-parent/tool-schemas.1.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -371,7 +371,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -389,15 +389,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-structured-parent/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/subagent-tool-filter/tool-schemas.1.expected.json

@@ -176,17 +176,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -270,7 +270,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -288,15 +288,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/system-prompt-in-history/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/team-targets/tool-schemas.expected.json

@@ -253,17 +253,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }

snapshots/session/text-turn/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/web-fetch/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/workflow-confinement/tool-schemas.expected.json

@@ -262,17 +262,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -356,7 +356,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -374,15 +374,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/workspace-dependencies/tool-schemas.expected.json

@@ -217,17 +217,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -319,7 +319,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -337,15 +337,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/session/worktree/tool-schemas.expected.json

@@ -279,17 +279,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -373,7 +373,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -391,15 +391,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/web/fresh-round-trip/tool-schemas.expected.json

@@ -325,17 +325,17 @@
     },
     {
       "name": "list_subagent_models",
-      "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.",
+      "description": "Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation.",
       "parameters": {
         "type": "object",
         "properties": {
           "provider": {
             "type": "string",
-            "description": "Registered LLM provider id. Omit to list providers."
+            "description": "Registered LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Exact model id to inspect. Requires provider; omit to list that provider's allowed models."
+            "description": "Exact model id; requires provider."
           }
         }
       }
@@ -719,7 +719,7 @@
     },
     {
       "name": "subagent",
-      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.",
+      "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default.",
       "parameters": {
         "type": "object",
         "properties": {
@@ -737,15 +737,15 @@
           },
           "provider": {
             "type": "string",
-            "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."
+            "description": "Child LLM provider id."
           },
           "model": {
             "type": "string",
-            "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."
+            "description": "Model id within that provider."
           },
           "reasoning_effort": {
             "type": "string",
-            "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."
+            "description": "Model-specific reasoning effort id."
           }
         },
         "required": [

snapshots/web/official-subagent-bundle/system-prompt.expected.md

@@ -158,11 +158,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Declare existing files as final deliverables for the user. Use it when the user needs a separate file, especially Office documents, spreadsheets, and slide decks; prefer your final response when that suffices. The user opens the current files; their contents are not copied. */
@@ -288,7 +288,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -296,11 +296,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The completion notice includes its final answer. This backend does not accept follow-up messages. */

snapshots/web/optional-string-editor/system-prompt.expected.md

@@ -158,11 +158,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Declare existing files as final deliverables for the user. Use it when the user needs a separate file, especially Office documents, spreadsheets, and slide decks; prefer your final response when that suffices. The user opens the current files; their contents are not copied. */
@@ -305,7 +305,7 @@ interface ToolArgsMap {
     /** Optional parameter of `view` command when `path` points to a file. If omitted or null, the full file is shown. If provided, the file will be shown in the indicated line number range, e.g. [11, 12] will show lines 11 and 12. Indexing at 1 to start. Setting `[start_line, -1]` shows all lines from `start_line` to the end of the file. */
     view_range?: number[] | null;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -313,11 +313,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */

snapshots/web/ptc-round/system-prompt.expected.md

@@ -158,11 +158,11 @@ interface ToolArgsMap {
     /** children (default) lists direct children, which accept send_message in any status. descendants lists the whole tree below you with each entry's parent session id and depth; entries deeper than 1 accept only interrupt_agent. */
     scope?: "children" | "descendants";
   } & Record<string, JsonValue>;
-  /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its allowed models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */
+  /** Inspect subagent models without changing the current Agent. No arguments: providers; `provider`: allowed models; `provider` + `model`: that model and its reasoning efforts. Use returned ids for delegation. */
   list_subagent_models: {
-    /** Registered LLM provider id. Omit to list providers. */
+    /** Registered LLM provider id. */
     provider?: string;
-    /** Exact model id to inspect. Requires provider; omit to list that provider's allowed models. */
+    /** Exact model id; requires provider. */
     model?: string;
   } & Record<string, JsonValue>;
   /** Declare existing files as final deliverables for the user. Use it when the user needs a separate file, especially Office documents, spreadsheets, and slide decks; prefer your final response when that suffices. The user opens the current files; their contents are not copied. */
@@ -288,7 +288,7 @@ interface ToolArgsMap {
     /** The exact skill name from the available skills list. */
     name: string;
   } & Record<string, JsonValue>;
-  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */
+  /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. Inspect options with `list_subagent_models` before setting `provider` and `model` together. Omitted values use child defaults, falling back to compatible parent values. When the effective route changes, an omitted `reasoning_effort` uses the model's default. */
   subagent: {
     /** Initial child working directory. Relative paths use your current directory; omitted inherits it. Later directory changes in either agent are independent. */
     cwd?: string;
@@ -296,11 +296,11 @@ interface ToolArgsMap {
     description: string;
     /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */
     prompt: string;
-    /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */
+    /** Child LLM provider id. */
     provider?: string;
-    /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */
+    /** Model id within that provider. */
     model?: string;
-    /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */
+    /** Model-specific reasoning effort id. */
     reasoning_effort?: string;
   } & Record<string, JsonValue>;
   /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This tool starts an independently managed subagent and immediately returns its id. The runtime notifies you when it finishes. The child reports results with `send_message`; use `send_message` to steer it while running or continue its conversation after it finishes. */