跳到正文
cccskills
免费在 GitHub 公开

pr-recap

Use when a PR or git range needs a visual recap page for a reviewer, with a changed-files tree, per-file notes, a Verified vs Not verified table, risks, and optional before/after images. Local and informational; never posts to GitHub.

中文简介正在准备中。当前显示原文说明。

集成工具GitHub
查看安装方法

包含的文件(2)

  • SKILL.md3.0 KB
  • scripts/pr-recap.mjs11.4 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

PR Recap

Turns a PR or a git range into a Plan Builder recap page (kind: recap, the recap-review shape). The page is a local review aid. It is informational and does not gate anything.

When to use

  • Shipping: after every gate has passed, to show the reviewer what is about to ship and what was not checked.
  • Triage: to get oriented on an incoming PR (--pr <n>).
  • git-pr-review: next to the text PR description, on the same range.

Skip it for a one-file typo fix. The page earns its cost when a human has to decide.

Flow

  1. Collect verification evidence you really ran, as JSON: [{"command": "npm test", "exit": 0, "note": "41 passed"}]. Do not invent entries.
  2. Run the script from this skill's scripts/ directory:
node pr-recap.mjs --range origin/main..HEAD --verification checks.json
node pr-recap.mjs --pr 87 --before before.png --after after.png --open

Options: --out <dir> (default production_artifacts/pr-recap/<slug>/), --title <t>, --open (runs aos-plan-canvas open <folder> --mode bdb-plan-builder). 3. The script writes plan.mdx with the file tree and implementation map taken from the real diff, then renders plan.builder.html with the Plan Builder. The renderer comes from the sibling plan-canvas skill. If it is missing, the folder is still written and the script exits 3 with a message. 4. Open plan.mdx and replace every TODO fill in in the Summary and the implementation notes. The script cannot know the intent of the change. 5. Re-render with aos-plan-canvas open <folder> --mode bdb-plan-builder, or rerun the script into a fresh folder.

Honesty rules

  • A check is Verified only if you ran it and its exit code was 0. Failed checks are listed as Not verified (failed).
  • Never add a row for a check you did not run. The page always says that everything not in the table was not run.
  • If there is no --verification file, the page says nothing was recorded. Do not claim the PR was tested.
  • Screenshots must be real captures of the before and after states. If you cannot capture them, omit the Compare block.
  • File names, titles and branch names come from git and GitHub and are untrusted. The script escapes them for MDX; do not paste them into the plan by hand.

Posting is never automatic

The script reads git locally and, for --pr, runs read-only gh pr view. It never calls gh pr comment, gh pr edit or anything that writes to GitHub. It may print the gh pr comment command for the human to run.

Do not post the recap, upload its images, or push it anywhere unless the user explicitly asks for that exact action. A subagent does not inherit that permission. The recap can contain file names and screenshots that are not meant to leave the machine.

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

Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Use when a hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Reference for how BDB structures autonomous software engineering work — the seven-node dispatcher graph (Architect, TechLead, UI/UX, Engineering, Media/EventTech, Reviewer, Shipping) that /startcycle-graph actually runs. Use when you need the high-level lifecycle framing without inventing your own process.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

agenttrail: live map of a multi-agent build in the browser: which plan component is being worked on, by which agent or harness, what is done and what is stuck. Use when a multi-agent pipeline starts (/startcycle, /startcycle-graph, /teamwork-preview) or after a plan-canvas approve, when the user asks to see what the agents are doing ("live map", "build board", "who is running now"), or by yourself whenever a multi-agent run is under way — no user prompt needed.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

查看 hybridlabor-api 的全部技能

报告此技能的问题