跳到正文
cccskills
免费在 GitHub 公开

executing-plans

Use when continuing implementation from an existing plan

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md2.1 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Executing Plans

Read the plan and current progress, then confirm its assumptions against the repository. Preserve existing decisions and completed work. Resolve gaps within the approved design; do not ask the user again after the design approval.

Use an existing isolated workspace or the project's workspace policy. The parent may execute tasks directly. Delegate independent bounded work only when the handoff and integration cost is worthwhile; tool availability alone is not a reason to switch workflows.

For each coherent task, implement the required behavior and run appropriate focused verification. Adjust mechanical steps when evidence warrants it while preserving scope and contracts. Track enough progress to survive interruption without reproducing tool logs by hand.

Investigate failures instead of merely stopping at the first failed command. Decide within the approved design; record deviations from the design and report them in the final report. Stop only for the exceptions in documents/development/development-policy.md §1.0 "承認後の進め方".

Complete the project's independent quality-check once for the final change, using its review budget and required verification. Do not add per-task reviews or another final code reviewer. Integrate according to existing authorization and project branch policy; use finishing-a-development-branch when integration guidance is needed.

Only the quality-check review limit (and a production's final quality review limit) stops the work (exception X1); a plan or mutation review limit is recorded and passed. If that limit is reached and the user approves exceeding it, run extend yourself with the same review-budget script used in step 1 of the quality-check reservation (.claude/hooks/…, .codex/hooks/…, or npx --no ai-dev-helm review-budget (where the CLI is not installed, npx -y @crearize/ai-dev-helm@<version in .ai-dev-helm.json> review-budget)): extend --phase <phase> --rounds <1-3> --reason "<the user's approval, quoted>"; never ask the user to run commands, inspect state, or reset state.

评价

还没有评价。欢迎分享你的使用感受。

评论

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。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

バックエンド実装時に使用。DRY原則遵守。コーディング規約準拠。

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

Use before implementing any feature, behavior change, or refactor - settles requirements and design, then gets the design independently reviewed and approved by the user before code is written

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

作業開始時に使用。mainブランチでの作業禁止。Issue先行作成必須。

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

UI実装後の検証時に使用。agent-browser CLIでブラウザ上の動作を手動検証する。「UIを確認」「画面テスト」と言われたら使用(プロジェクトの E2E スイート実行は quality-check Step 5(推奨度・範囲で自動実施または確認)/ server-startup が担当)。

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

DBマイグレーション作成時に使用。バージョン番号競合防止。mainブランチ確認必須。

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

Use when independent tasks benefit from parallel work without shared state or sequential dependencies

中文简介正在准备中。当前显示原文说明。

Crearize/ai-dev-helm52026年10月12日 更新

查看 Crearize 的全部技能

报告此技能的问题