Skip to content
Back to search

looks-good-to-ai

MarketplaceWorks with Claude CodeGitHub

An Agent Skill for reviewing AI-written pull requests: catch bugs, align context, own the sign-off.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/laulauraluo--ai-code-review-triage/badge.svg)](/listings/laulauraluo--ai-code-review-triage)

HTML

<a href="/listings/laulauraluo--ai-code-review-triage"><img src="/api/listings/laulauraluo--ai-code-review-triage/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

npx skills add laulauraluo/looks-good-to-ai

The [`skills`](https://github.com/vercel-labs/skills) CLI asks which agents to set up. Add `-g` to install for your user instead of this project, and `-a <agent>` to skip the question (`-a claude-code`, `-a codex`, `-a github-copilot`, `-a cursor`, …).

Or use the agent's own installer:

| Agent | Command | |---|---| | Claude Code | `/plugin marketplace add laulauraluo/looks-good-to-ai` then `/plugin install looks-good-to-ai@looks-good-to-ai` | | Codex | `codex plugin marketplace add laulauraluo/looks-good-to-ai` then `codex plugin add looks-good-to-ai@looks-good-to-ai` | | GitHub Copilot CLI | `copilot plugin marketplace add laulauraluo/looks-good-to-ai` then `copilot plugin install looks-good-to-ai@looks-good-to-ai` | | Gemini CLI | `gemini extensions install https://github.com/laulauraluo/looks-good-to-ai` |

Then, in your agent:

Review this AI-generated PR before I merge it.

The skill loads on its own from a request like that. Most agents also take `/looks-good-to-ai` to call it by name; Codex uses `$looks-good-to-ai`. [docs/agents.md](docs/agents.md) has the command for each agent.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 14 lines
  1. {
  2. "name": "looks-good-to-ai",
  3. "description": "An Agent Skill for reviewing AI-written pull requests: catch bugs, align context, own the sign-off.",
  4. "owner": {
  5. "name": "laulauraluo"
  6. },
  7. "plugins": [
  8. {
  9. "name": "looks-good-to-ai",
  10. "source": "./",
  11. "description": "Review AI-written PRs by splitting review into three jobs: catch bugs, align context, own the sign-off."
  12. }
  13. ]
  14. }
README.md · README · 293 lines
  1. <div align="center">
  2. # Looks Good to AI
  3. <sub>Formerly AI Code Review Triage</sub>
  4. **AI writes code faster than you can read it. Stop reading harder and review differently.**
  5. A code review [Agent Skill](https://agentskills.io) for AI-written pull requests. It splits "reading the code" into three separate jobs.<br>
  6. Runs in Claude Code, Codex, GitHub Copilot, Cursor, Gemini CLI, Antigravity, OpenCode, Windsurf, Cline, Roo Code, Kiro, Amp, Junie and Zed, and loads as plain context in Aider.
  7. <strong title="English" aria-label="English">🇬🇧</strong> ·
  8. <a href="README.zh-CN.md" title="简体中文" aria-label="简体中文">🇨🇳</a> ·
  9. <a href="README.ja.md" title="日本語" aria-label="日本語">🇯🇵</a> ·
  10. <a href="README.ko.md" title="한국어" aria-label="한국어">🇰🇷</a> ·
  11. <a href="README.es.md" title="Español" aria-label="Español">🇪🇸</a> ·
  12. <a href="README.pt-BR.md" title="Português (Brasil)" aria-label="Português (Brasil)">🇧🇷</a> ·
  13. <a href="README.fr.md" title="Français" aria-label="Français">🇫🇷</a>
  14. [![Agent Skills](https://img.shields.io/badge/Agent_Skills-compatible-1b1b19?style=flat-square)](https://agentskills.io/specification)
  15. [![Claude Code plugin](https://img.shields.io/badge/Claude_Code-plugin-d97757?style=flat-square)](#claude-code-plugin)
  16. [![License: MIT](https://img.shields.io/badge/license-MIT-1b1b19?style=flat-square)](LICENSE)
  17. [![Version](https://img.shields.io/badge/version-2.0.0-1b1b19?style=flat-square)](CHANGELOG.md)
  18. [![CI](https://github.com/laulauraluo/looks-good-to-ai/actions/workflows/ci.yml/badge.svg)](https://github.com/laulauraluo/looks-good-to-ai/actions/workflows/ci.yml)
  19. <a href="#quick-start">Quick start</a> ·
  20. <a href="#the-idea">The idea</a> ·
  21. <a href="#install">Install</a> ·
  22. <a href="#usage">Usage</a> ·
  23. <a href="#examples">Examples</a> ·
  24. <a href="#how-it-works">How it works</a> ·
  25. <a href="#limits">Limits</a>
  26. <img src="assets/before-after.svg" alt="An AI-written PR whose tests pass and whose diff looks clean. Looks Good to AI finds four problems: a dropped unit conversion hidden in the diff, a past revert, a mobile client that hard-codes the old value, and a requirement mismatch. Read line by line, it logs every user out after 30 seconds; with the skill, it is blocked before merge" width="820">
  27. </div>
  28. ---
  29. ## The problem
  30. Agent-written PRs pile up faster than anyone can review them line by line. Reading harder doesn't help much either: the code is tidy on the surface, so the real problems hide well enough to survive a careful read. "Just merge and roll back if it breaks" works technically, but a human still answers for the incident.
  31. ## Quick start
  32. ```sh
  33. npx skills add laulauraluo/looks-good-to-ai
  34. ```
  35. The [`skills`](https://github.com/vercel-labs/skills) CLI asks which agents to set up. Add `-g` to install for your user instead of this project, and `-a <agent>` to skip the question (`-a claude-code`, `-a codex`, `-a github-copilot`, `-a cursor`, …).
  36. Or use the agent's own installer:
  37. | Agent | Command |
  38. |---|---|
  39. | Claude Code | `/plugin marketplace add laulauraluo/looks-good-to-ai` then `/plugin install looks-good-to-ai@looks-good-to-ai` |
  40. | Codex | `codex plugin marketplace add laulauraluo/looks-good-to-ai` then `codex plugin add looks-good-to-ai@looks-good-to-ai` |
  41. | GitHub Copilot CLI | `copilot plugin marketplace add laulauraluo/looks-good-to-ai` then `copilot plugin install looks-good-to-ai@looks-good-to-ai` |
  42. | Gemini CLI | `gemini extensions install https://github.com/laulauraluo/looks-good-to-ai` |
  43. Then, in your agent:
  44. ```text
  45. Review this AI-generated PR before I merge it.
  46. ```
  47. The skill loads on its own from a request like that. Most agents also take `/looks-good-to-ai` to call it by name; Codex uses `$looks-good-to-ai`. [docs/agents.md](docs/agents.md) has the command for each agent.
  48. ## The idea
  49. "Reviewing code" has always bundled three different jobs. That worked when code arrived slowly. At agent speed, each job needs its own process:
  50. ```mermaid
  51. flowchart LR
  52. PR([AI-written PR]) --> T{Step 0<br/>Triage<br/>risk tier}
  53. T --> B["**1 · Catch bugs**<br/>fresh-context sub-agent<br/>+ tests / repro"]
  54. B --> C["**2 · Align context**<br/>orders of magnitude<br/>history & neighbors<br/>requirement match"]
  55. C --> S["**3 · Own the sign-off**<br/>explain · detect · undo<br/>scoped approval"]
  56. S --> R([Review report])
  57. ```
  58. | Job | Question | Who does it best |
  59. |---|---|---|
  60. | **Catch bugs** | Is the code wrong? | Machines, with an *independent* second pass |
  61. | **Align context** | Is this the right change for this product, today? | A human, checking angles the diff can't show |
  62. | **Own the sign-off** | Will I put my name on this? | The approver, with an explicit scope |
  63. Low-risk PRs only run Step 1. Medium-risk PRs run Steps 1 and 2. High-risk PRs (production behavior, data, security, money) run all three.
  64. ### What it catches that line-by-line review misses
  65. > A PR changes `SESSION_TIMEOUT = 30` to `30000` and drops one line of unit conversion. Tests pass. The code is clean. A line-by-line reader approves it in 30 seconds.
  66. >
  67. > - **Step 1** has a fresh-context sub-agent add a boundary test. Result: sessions now expire after **30 seconds**.
  68. > - **Step 2** checks history and callers. The same change was reverted once before, the mobile client hard-codes the old value, and the request was only for "remember me" users while the diff changes everyone.
  69. > - **Step 3** blocks the approval. Nobody can explain what the PR actually does.
  70. >
  71. > Full run: [examples/session-timeout-walkthrough.md](examples/session-timeout-walkthrough.md)
  72. ## Install
  73. The skill is the folder [`skills/looks-good-to-ai/`](skills/looks-good-to-ai). Every agent below loads it unchanged; only the folder differs. [docs/agents.md](docs/agents.md) has the full table with every folder each agent reads, the way each agent calls a skill, and a link to each agent's own documentation.
  74. | Agent | skills CLI | Or copy the folder to (user · project) |
  75. |---|---|---|
  76. | Claude Code | `-a claude-code` | `~/.claude/skills/` · `.claude/skills/` |
  77. | Codex | `-a codex` | `~/.agents/skills/` · `.agents/skills/` |
  78. | GitHub Copilot | `-a github-copilot` | `~/.copilot/skills/` · `.github/skills/` |
  79. | Cursor | `-a cursor` | `~/.cursor/skills/` · `.cursor/skills/` |
  80. | Gemini CLI | `-a gemini-cli` | `~/.gemini/skills/` · `.gemini/skills/` |
  81. | Antigravity | `-a antigravity` | `~/.gemini/config/skills/` · `.agents/skills/` |
  82. | OpenCode | `-a opencode` | `~/.config/opencode/skills/` · `.opencode/skills/` |
  83. | Windsurf | `-a windsurf` | `~/.config/devin/skills/` · `.devin/skills/` |
  84. | Cline | copy by hand ([why](docs/agents.md#cline)) | `~/.cline/skills/` · `.cline/skills/` |
  85. | Roo Code | `-a roo` | `~/.roo/skills/` · `.roo/skills/` |
  86. | Kiro | `-a kiro-cli` | `~/.kiro/skills/` · `.kiro/skills/` |
  87. | Amp | `-a amp` | `~/.config/agents/skills/` · `.agents/skills/` |
  88. | Junie | `-a junie` | `~/.junie/skills/` · `.junie/skills/` |
  89. | Zed | `-a zed` | `~/.agents/skills/` · `.agents/skills/` |
  90. | Aider | — | `read:` in `.aider.conf.yml` ([how](docs/agents.md#aider)) |
  91. | Continue | — | Unverified; load as a rule ([how](docs/agents.md#continue)) |
  92. <details>
  93. <summary><b>With git, no CLI</b></summary>
  94. <br>
  95. Clone the repository once and copy the skill folder into the agent's folder from the table, so that `SKILL.md` ends up at `<folder>/looks-good-to-ai/SKILL.md`:
  96. ```sh
  97. git clone https://github.com/laulauraluo/looks-good-to-ai.git
  98. cp -r looks-good-to-ai/skills/looks-good-to-ai ~/.claude/skills/
  99. ```
  100. Swap `~/.claude/skills/` for any other folder in the table. For a project install, commit the copied folder so your team gets it too.
  101. </details>
  102. <details>
  103. <summary><b>On every pull request (GitHub Actions)</b></summary>
  104. <br>
  105. A reusable workflow runs the skill on each pull request and posts the report as a comment. Setup takes three steps: [docs/github-action.md](docs/github-action.md).
  106. </details>
  107. <details id="claude-code-plugin">
  108. <summary><b>Claude Code plugin</b></summary>
  109. <br>
  110. Inside Claude Code:
  111. ```text
  112. /plugin marketplace add laulauraluo/looks-good-to-ai
  113. /plugin install looks-good-to-ai@looks-good-to-ai
  114. ```
  115. </details>
  116. ## Usage
  117. Ask your agent something like:
  118. - *"Review this AI-generated PR before I merge it."*
  119. - *"Can this diff be merged? Codex wrote it."*
  120. - *"I'm about to approve #123. What should I verify first?"*
  121. - *"We have too many agent PRs to read line by line. Triage this one."*
  122. You get a report like this (from the invented timeout example):
  123. ```markdown
  124. ## Extend session lifetime for "remember me" · Risk: High
  125. ### 1. Bugs (independent fresh-context review)
  126. - [CONFIRMED] auth/session.py:3 — unit conversion dropped; sessions expire after 30 s. Repro: test_session_expires_after_timeout
  127. ### 2. Context
  128. - Magnitudes: SESSION_TIMEOUT 30 → 30000 (comment says min, code treats as ms) — suspicious
  129. - History / neighbors: reverted once before (a1b2c3d); mobile/src/config.ts hard-codes 30 min
  130. - Requirement match: asked for remember-me users only; diff changes everyone
  131. ### 3. Sign-off
  132. - Detection: login-rate and 401-rate alerts · Rollback: config revert
  133. - Recommendation: request changes
  134. - Verification scope: verified units and callers; did not load-test
  135. <sub>Reviewed with [Looks Good to AI](https://github.com/laulauraluo/looks-good-to-ai)</sub>
  136. ```
  137. The last line credits the skill when you paste the report into a PR. Ask the agent to leave it out if you don't want it.
  138. <details>
  139. <summary><b>More prompts</b>, from the skill's <a href="evals/evals.json">evals</a></summary>
  140. <br>
  141. ```text
  142. Here's the diff for PR #482. Cursor wrote it to raise our retry count from 3 to 10 on the payments client. Tests pass. Can I approve it?
  143. ```
  144. ```text
  145. Copilot opened a PR that only fixes typos in docs/ and the README. Do I need the full review process?
  146. ```
  147. ```text
  148. I wrote this change with you in this session. Now review it before I merge.
  149. ```
  150. </details>
  151. ## Examples
  152. One real PR, reviewed after the fact, and two invented walkthroughs that follow one PR through every step.
  153. | Walkthrough | What goes wrong | What finds it |
  154. |---|---|---|
  155. | [Real PR: LightBulb #413](examples/lightbulb-real-pr.md) | A Copilot PR to cut idle CPU/RAM, merged after dozens of review rounds with full coverage, reverted eight months later. | A blind Step 1 pass flagged the Settings bug users later reported; Step 2 found that nobody had measured CPU or RAM, the reason for the revert. |
  156. | [Session timeout](examples/session-timeout-walkthrough.md) | `30` → `30000` with the unit conversion dropped: sessions expire after 30 seconds, for every user. | A boundary test in Step 1; units, a past revert and a hard-coded mobile value in Step 2. |
  157. | [Pagination limit](examples/pagination-limit-walkthrough.md) | Page size `100` → `1000` against a provider that caps pages at 100: the export stops with 150 of 250 rows missing. | The provider contract in Step 2, and a standard-library Python reproduction you can run. |
  158. ## How it works
  159. - **[`SKILL.md`](skills/looks-good-to-ai/SKILL.md)** is what the agent reads: the risk tiers, the three steps, the commands for history and callers, and the report template.
  160. - **Step 1 needs a second, clean context.** The session that wrote the code can't review it independently. [`references/subagents.md`](skills/looks-good-to-ai/references/subagents.md) shows how each agent starts a fresh reviewer, and what to do in agents that have none.
  161. - **Step 2 needs a shell with `git`.** It runs `git log -S`, `git log --grep=revert` and `grep` to find history and callers the diff doesn't show.
  162. - **[`evals/evals.json`](evals/evals.json)** holds test prompts and what a good answer contains. CI checks that the skill's frontmatter follows the [Agent Skills spec](https://agentskills.io/specification) and that the pagination reproduction still runs.
  163. <details>
  164. <summary><b>Repository layout</b></summary>
  165. <br>
  166. ```text
  167. looks-good-to-ai/
  168. ├── skills/looks-good-to-ai/ the skill (this folder is what gets installed)
  169. │ ├── SKILL.md workflow, risk tiers, report template
  170. │ ├── references/subagents.md a fresh-context reviewer in each agent
  171. │ └── agents/openai.yaml display name for the Codex and ChatGPT apps
  172. ├── docs/
  173. │ ├── agents.md install and use in each agent, with sources
  174. │ └── review-ai-generated-code.md questions and answers
  175. ├── examples/ one real PR and two invented walkthroughs
  176. ├── evals/evals.json test prompts and expected answers
  177. ├── scripts/check.py the checks CI runs
  178. ├── .claude-plugin/ plugin and marketplace manifests (Claude Code, Codex, Copilot CLI)
  179. ├── gemini-extension.json Gemini CLI extension manifest
  180. └── .github/ CI, issue forms, PR template
  181. ```
  182. </details>
  183. ## Limits
  184. - **It is a review process, not a scanner.** It decides what has to be true before you approve. Run linters, type checkers and security tools inside Step 1; this skill does not replace them.
  185. - **It can't guarantee a PR is safe.** The real PR was picked because it was already reverted, and the other two examples are invented. None of them is a measured detection rate.
  186. - **It is slower than a skim, on purpose.** A high-risk PR gets a sub-agent pass, tests, history searches and a written sign-off. Low-risk PRs skip most of it.
  187. - **Step 1 is only as independent as the reviewer you give it.** In agents without sub-agents, open a new session by hand.
  188. Questions such as "can it review Codex output?" and "why can passing tests miss incomplete exports?" are answered in [docs/review-ai-generated-code.md](docs/review-ai-generated-code.md).
  189. ## Works well with
  190. This skill organizes the review and decides what has to be true before you approve. It doesn't replace specialized checkers. Run them inside Step 1:
  191. - a correctness-focused review (for example Claude Code's built-in `/code-review`, or Copilot CLI's code-review agent)
  192. - security and performance checklists
  193. - over-engineering reviews
  194. ## Contributing
  195. Issues and pull requests are welcome. Read [CONTRIBUTING.md](CONTRIBUTING.md) first: it covers linking a working copy into your agent, the checks CI runs, and what a change to `SKILL.md` has to keep.
  196. - [Report a problem](https://github.com/laulauraluo/looks-good-to-ai/issues/new?template=bug_report.yml)
  197. - [Add or fix an agent](https://github.com/laulauraluo/looks-good-to-ai/issues/new?template=agent_support.yml)
  198. - [Changelog](CHANGELOG.md)
  199. ## Acknowledgments
  200. The three-job framing was inspired by [a public talk](https://www.bilibili.com/video/BV1mRn66JERc/) on reviewing AI-written code. The multi-agent install layout follows [Anatomy](https://github.com/wheresryan22/anatomy).
  201. ## License
  202. [MIT](LICENSE)
  203. ---
  204. <div align="center">
  205. If this caught a bad merge for you, a ⭐ helps other reviewers find it.
  206. </div>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.