Design commands and processing loops that reach the same final state after retries or crashes. Use the current state to apply only necessary changes, and make leftover locks, files, or records recognizable during the next run.
中文简介正在准备中。当前显示英语简介。
Apply when designing commands, lifecycle steps, or processing loops that run amid crashes, restarts, and retries. Converge to the same end state regardless of partial prior runs.
安装前可以查看交给 AI 代理的指令内容。
Design operations so they converge to the correct state regardless of how many times they run or where the previous run stopped. Every state-mutating operation should answer: "What happens if this runs twice? What happens if the previous run crashed halfway?"
Why: Commands, lifecycle operations, and processing loops run where crashes, restarts, and retries are normal. If partial state changes the next run's outcome, every restart becomes a debugging session.
The idea: derive each step from the state that exists, not from what the previous run meant to do. Read the current state, compare it to the goal, and apply only the difference. A step written this way is safe to repeat, because a repeat finds nothing left to do.
Anything a run leaves behind, such as a lock, a temporary file, a half-written record, or a claimed job, must be recognizable as stale by the next run. Judge staleness by evidence that stays true wherever the previous run could have stopped, not only in the common case. A marker that the next run cannot judge turns every crash into manual cleanup. For example, write to a temporary path and rename it into place, or compare artifacts by content instead of by creation order.
The test:
If any answer is "it depends on what state was left behind," the operation needs a reconciliation step.
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。
Plan types, function inputs and outputs, and module boundaries before implementation. Compare structurally different designs from multiple AI models, document the chosen approach, and guide coding while tracking changes to the agreed design.
中文简介正在准备中。当前显示英语简介。
Generate several candidate solutions in parallel, compare them against concrete criteria, and choose a base. Combine useful parts from other candidates, then verify the final artifact and document why each major choice was made.
中文简介正在准备中。当前显示英语简介。
Turn recurring work habits and communication preferences into a personal agent skill. It combines evidence from workspace conversations with direct questions, then drafts or updates instructions for how agents should work with you.
中文简介正在准备中。当前显示英语简介。
Monitor an open pull request, fix failing automated checks and merge conflicts, and address straightforward review feedback. Recheck after updates, report the resulting status, and hand decisions requiring design judgment back to the user.
中文简介正在准备中。当前显示英语简介。
Checks whether performance measurements support a speedup, slowdown, or technology choice. It examines bottlenecks, production settings, errors, repeatability, and completed work, then guides reporting with run counts, ranges, and a clear verdict.
中文简介正在准备中。当前显示英语简介。
Investigate what a code change could break beyond the edited files. Trace downstream effects, identify the key assumption behind its safety, and test that assumption with real code while clearly reporting unresolved risks.
中文简介正在准备中。当前显示英语简介。