本文へ移動
cccskills
無料GitHub で公開

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.

インストール方法を見る

含まれるファイル(2)

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

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

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.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

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

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12日 更新

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

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12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.

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12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.

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12日 更新

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

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12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.

日本語の概要は準備中です。原文の説明を表示しています。

hybridlabor-api/aos62026年10月12日 更新

hybridlabor-api のスキルをすべて見る

このスキルの問題を報告する