Give durable user context its producer-owned source kind instead of a generic plugin wrapper. The V3-to-V4 conversion preserves the frozen rename and collision rules, all own JSON metadata, event coordinates, and historical request-only kinds. Native catalog and JSONL admission retain source validation so unknown attribution can survive without its producer while malformed owned sources are refused. Use an explicit user-only RequestUserInput for auxiliary LLM prompts that do not own durable message identity. Auto Review and compaction remove their request-only source registrations while retaining prompt content and freezing; provider serializers and projections keep the same wire data, assistant replay isolation, image ordering, and tool correlation. Session writes, Agent delivery, and recorded title requests continue to require identified durable messages. Bind the saved compatibility policy only to user sources and qualify tmux location attribution without weakening fixed system, model, or tool semantics. Rebuild the current inventory and unpublished V4 acknowledgement on the actual tool-role parent, preserving its rationale and all accepted historical records. The catalog format version records tooling metadata; Session remains V4. Developer-role policy and activation belong to their own feature. The latest-master Messages adapter, serialization, request-input typing, EpochHeader reservation, and attribution checks pass 103 tests across five files. The retained V3 Messages input-history snapshot passes through the source profile; Host/Client typechecks and all 41 documentation gates pass. The generated comparison records four changed roots and 447 unchanged type fingerprints. Latest-master ARIA changes are retained, with producer source labels applied independently; new Messages fixtures retain their producer names as declared source kinds and use the fixed system-message helper; recorded Session generations are unchanged by this regrouping. Co-authored-by: Tianyi Cui <[email protected]>
description, kind
| description | kind |
|---|---|
| Package map for the loop-hygiene guard family: the advisory repeat-tool reminder and the per-tool-call timeout policy, for users and maintainers choosing or composing the guards. | package-group |
guard/ — loop-hygiene guard family
English | 中文
Summary
The guard/ group keeps the agent loop productive by watching for two common failure patterns. repeat-tool-reminder notices when the model repeats the exact same tool call and reminds it to change approach or finish, so a stuck loop stops burning time and tokens. timeout-policy puts a time limit on tool calls that declare one, so a hung call returns a clear timed-out error to the model instead of stalling the session. Both ship enabled in the dsh base bundle; a composition can tune or remove them.
Table of Contents
Packages
Two small plugins cover the two patterns; each README below explains when to keep, tune, or remove it.
| Package | What it provides |
|---|---|
repeat-tool-reminder/ |
Reminds the model when it repeats the same tool call, so it changes approach or finishes |
timeout-policy/ |
Times out tool calls that declare a limit, so the model gets a clear error instead of waiting forever |
Related documentation
Start with the tools subsystem reference for the tool-call pipeline, then the reminder's configuration and the timeout-library decision behind the policy.
- Tools subsystem reference — the tool-call pipeline and decisions both guards build on.
- Generated configuration catalog — every accepted field of the repeat-call reminder.
- Timeout deadline library Agent Note — the timing/termination split
timeout-policyenforces.
Dev Note
Working context for maintainers — click to expand
None.