agentleFS
mannobeatsmannobeats/hyper-taste/skills/review/SKILL.md

review

Skill1 starsChanged 4 days agoMITView on GitHub

Review code before it merges, or audit a whole codebase - correctness, security, data safety, performance, tests, duplication and architecture - and report findings with a concrete failing case and the smallest fix. Use when asked to review a diff, PR, branch or file, audit a repo, check code an agent wrote, or at the end of a milestone.

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, 4 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 157 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Scale: check-then-write races, a query per item, lists or memory that only grow, work repeated by every process, O(n²) on real input.

    line 28
  2. Distinctive:

    Every finding has a concrete case: "this input or situation leads to this wrong result".

    line 42
  3. Distinctive:

    Before calling code unused or duplicated, grep the whole tree, including tests, config and dynamic references.

    line 44
  4. Distinctive:

    Every changed line was read, every finding has a case and a fix, and the report says what was not checked.

    line 64

“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

The outcome: the code that merges is correct, safe and part of one coherent system, and every finding comes with proof. A review that lists taste and vague worries teaches nothing.

1. Map first

  • What is it supposed to do? Read the story, the brief or the PR description, then the diff.
  • For an audit: the README, entry points, config, tests, and the expected load (one person or many users and processes at once). Go deep where a mistake costs the most (input, money, auth, data writes, background jobs), and say which parts you did not read.
  • Trace the main flows end to end: where data comes in, what is stored, what goes out.

2. Look for, in this order

  1. Correctness: wrong results, crashes, missed edge cases (empty, zero, last item, rounding, time zones, concurrent edits), callers that disagree with what a function returns, the same rule applied differently in two places, acceptance criteria not met.
  2. Security: unvalidated input at trust boundaries, injection, broken auth or access checks, secrets in code or logs, unsafe defaults, data exposed to the wrong user, risky dependencies.
  3. Data safety: swallowed errors, writes in the wrong order, missing transactions, migrations that lose data, no way to undo a destructive action.
  4. Scale: check-then-write races, a query per item, lists or memory that only grow, work repeated by every process, O(n²) on real input. Judge against the expected load.
  5. Tests: risky logic with no test that fails when it breaks; tests that pass without testing anything; skipped or loosened tests.
  6. One system: a second version of a helper or component that already exists, near-copies that must change together, a rule that lives in two places, an abstraction with one user, code that ignores the project's conventions, tokens or design system.
  7. Architecture: a module doing several unrelated jobs, logic in the wrong layer, a change that has to touch many files for one feature. Propose a deeper module only with a clear win.
  8. Performance: measured problems first. Guesses are suggestions, not findings.
  9. Readability: names that mislead, comments that lie. No style taste.

3. Prove each finding

  • Every finding has a concrete case: "this input or situation leads to this wrong result". No case, no finding.
  • Before calling code unused or duplicated, grep the whole tree, including tests, config and dynamic references.
  • A shortcut with a comment naming its limit is a decision, not a finding, unless the load already crosses it.
  • Propose the smallest fix that solves it, preferring fixes that remove code.

4. Report

Plain words, worst first, in three groups (skip empty ones):

  • Must fix: bugs, security, data loss, breaks at the expected load.
  • Should fix: risky code without a test, duplication, real slowness, wrong layer.
  • Could fix: small, clear improvements.

Each line: file:line, what is wrong, the case that shows it, the fix. End with what you did not check. If the code is good, say so and say what you checked. Fixing a large set of findings (dead code, duplicates, structure) is refactor.

Done when

Every changed line was read, every finding has a case and a fix, and the report says what was not checked. When fixing your own findings, each fix gets a test or a look at the running result.

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.