Use when the user needs a standup-ready summary of recent project activity from local evidence.
中文简介正在准备中。当前显示原文说明。
Use when verification depends on agent runtime phases, queued input, tool-call lifecycle, save points, retry, abort, compaction, or branch behavior.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Lifecycle stage: VERIFY
Use when an artifact, adapter, handoff, runtime smoke run, scheduled job, or long-running goal claims behavior across turn boundaries, queued messages, tool execution, compaction, retry, abort, branch switching, or session persistence.
Do not use for ordinary artifact checks where the claim is fully proven by static files, schemas, tests, or docs and no runtime phase, queue, tool, or persistence boundary affects trust.
| Shortcut | Rebuttal |
|---|---|
| "The transcript shows the message, so it was applied." | A transcript line is not proof of delivery timing. Name the queue and delivery boundary. |
| "The setting changed, so the current turn used it." | In-flight snapshots may not change. Prove whether the change affected this turn or the next one. |
| "The tool finished first, so its result is first." | Parallel completion order can differ from source-order persistence. Check final durable order. |
| "Abort means nothing was written." | Abort can leave pending writes or partial evidence. Check save-point and cleanup evidence. |
| "Compaction preserved enough." | Compaction must name what was summarized, kept, omitted, and when it becomes stale. |
Use templates/qa-evidence.md for verification claims or templates/handoff-report.md for blocked resumes. Include evidence, blockers, next action, phase, queued input, tool lifecycle evidence, save point, read files, modified files, and stale-proof risk.
Trigger: a runtime smoke transcript claims a follow-up message changed tools during an active verification run. Action: classify the phase, identify whether the input was steering or follow-up, inspect tool preflight and persisted result order, and block if the transcript lacks save-point proof. Output artifact: templates/qa-evidence.md with blockers and next action. Verification: cite the transcript path, runtime smoke artifact, selected command, loaded skill file, queue evidence, tool lifecycle evidence, and validation command.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Use when the user needs a standup-ready summary of recent project activity from local evidence.
中文简介正在准备中。当前显示原文说明。
Use when a command, adapter, or runtime smoke depends on proving an agent runtime's capabilities before trusting command routing, writes, shell access, or validation claims.
中文简介正在准备中。当前显示原文说明。
Use when agent-produced work needs a safety, evidence, and reliability check before handoff, merge, or trust.
中文简介正在准备中。当前显示原文说明。
Use when an output artifact, template, schema, example, or handoff field must stay aligned across commands, skills, docs, and validators.
中文简介正在准备中。当前显示原文说明。
Agent Brain /brain-brief: after intake, research, and grill have enough signal.
中文简介正在准备中。当前显示原文说明。
Agent Brain /brain-build: an implementation plan has a selected task and validation method.
中文简介正在准备中。当前显示原文说明。