agentleFS
Sign inStart free
ArcticFox2029ArcticFox2029/chamnan/skills/review/SKILL.md

review

Skill10 starsChanged 8 days agoMITView on GitHub

Review a change with what this repository knows about it — who depends on the file, which tests cover it, what broke here before, and what the repo has written down. Use before approving a diff, opening a PR, or when asked to review.

192 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use close-spec, design-with-images and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • dzhngSkill

    close-spec

    When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).

    Also in 5 projects that use this

  • dzhngSkill

    design-with-images

    Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.

    Also in 5 projects that use this

  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文，过滤相关性，auto-ingest 高优先级论文，检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门：Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • +3 more from dzhng
  • +5 more from Lambenthan

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 157 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Size first: If the change is more than about 400 changed lines, say so before anything else and review it in parts, one coherent group of …

    line 18
  2. Very specific:

    3 · What broke here before: Look for a 🐛 marker on or near the changed lines.

    line 45
  3. Distinctive:

    If a rule, a dead end or a recorded lesson covers this, the review's job is to point at it — not to re-argue it.

    line 56
  4. Distinctive:

    Do not manufacture severity: If the four signals come back quiet, the change is quiet.

    line 74

“Common” is measured: kept by at least 6 in 10 of the 157 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Review it against the repository, not just against the diff

A diff shows what changed. It cannot show what depends on it, what broke here last time, or what this repository has already decided about it. Those are the four things worth saying, and three of them are one command.

chamnan-impact <each changed file>

That answers who imports it, which tests cover it, and which threads name it. Run it per changed file, not once for the change.

Size first. If the change is more than about 400 changed lines, say so before anything else and review it in parts, one coherent group of files at a time. Defect detection measured at 87% for reviews of 1-100 lines falls to 28% above 1,000 — a review of a large change finds less, and says it found everything. The same goes for a change that mixes a rename or move with a change in behaviour: say so, and review the behaviour on its own — reviewers of a tangled change reported six false positives against one for the same change split in two.

<!-- 🎯 [2026-09-24] (R47, 2026-09-24) The tangled-change figure is that round's ninth finding: 28 developers, one feature-plus-refactoring pair, p = 0.03. --><!-- 🎯 [2026-09-24] (R35, 2026-09-24) Screened from a research round on code review: the 87%-to-28% figure is its first finding, and the review skill had no word about size. -->

Read the change before its description. Work out what the diff does from the code and the four signals below first; only then read the commit message or pull-request text, and compare. A claim the code does not bear out, or a change the description never mentions, is a finding. Read the other way round, the description tells you what to see and you see it.

<!-- 🎯 [2026-10-07] (BMAD-METHOD, read 2026-10-07) bmad-review stages the change's own narrative for one lens only, so the others judge the code unprimed. Owner: "2 3 ดูดี". -->

The four signals, and what each one is for

1 · Who breaks. The import graph. Say the number and name the surprising ones — "six call sites, and two are in hooks that run on every tool call" is a review comment. "This file is imported" is not.

2 · Which tests cover it. From the same output. If a changed file has none, that is the finding — say so plainly, once, without turning it into a lecture about coverage.

3 · What broke here before. Look for a 🐛 marker on or near the changed lines. This project writes the defect record at the line it was fixed on, so a diff that touches one is touching a line somebody already got wrong. A change that reverts a 🐛 line is the single most valuable thing a review can catch, and no diff-only reviewer can see it.

4 · What the repository has already decided.

chamnan-recall <the subject of the change>

If a rule, a dead end or a recorded lesson covers this, the review's job is to point at it — not to re-argue it.

When the workspace is new

A fresh workspace has no 🐛 records and no recorded rules. Report the signals you have and stop. Two of four is still two more than a diff gives, and inventing the other two is worse than missing them.

What not to say

Do not restate the diff. The reader has it. Every line of the review should be something they could not have got by reading the change themselves — which is exactly the four signals above.

Do not report a stale index as fact. chamnan-impact says when the map is behind the code and tells you to run chamnan-map. That warning is part of the answer; passing its conclusions on without it turns "nothing references this" into a false all-clear.

Do not manufacture severity. If the four signals come back quiet, the change is quiet. Saying so in one line is a complete review.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free