agentleFS
jisuzzzzjisuzzzz/haechi/skills/review/SKILL.md

review

Skill17 starsChanged 14 days agoMITView on GitHub

Review — judge a diff or an area of existing code without changing it, using review-diff for the evidence standard and independent reviewers when useful. Use for /haechi:review, when the user asks to review a merge request, a branch, uncommitted changes or an area of code.

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

3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 157 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Set the focus: Default: logic defects, excess (code that does more than needed), sensitive data and safety.

    line 22
  2. Very specific:

    Only when the user asks. Copy the repo to a temporary folder outside the repository and run /haechi:mutation-check there with --base <base> …

    line 31
  3. Very specific:

    Every defect has file:line and a scenario (input or state, then the wrong result).

    line 35

“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 requested code

Apply the common work contract within the authorized scope. It owns required implementation, verification, review and completion obligations; this skill supplies methods. Use /haechi:delegate for a bounded worker task when independent execution helps. Existing step names below are compatible delegation entry points, not an exclusive workflow state.

The main session runs these steps. Scope, delegation, authorization and completion requirements are in the session rules. How to judge — the order to look in, the three buckets, the evidence each needs, the report shape — is all in /haechi:review-diff, so that a main-session review, a verifier review and a user-requested review use the same bar. This file adds only what is specific to a review request. Principle files are two levels above this skill's folder (<this skill's folder>/../../principles/<name>.md): evidence-levels and behavior-tests apply.

Task focus

Review without code edits. The easy mistake is fixing what you find during the review. Example: "Review src/invoices and report findings; the user decides what to fix."

Steps

  1. Pick the mode.
    • Diff review (default): a merge request, a branch against its base, or uncommitted changes.
    • Area review: existing code with no diff, for example "review src/orders". Inspect the target and relevant consumers for defects, risks and evidence-backed quality improvements. Separate improvement suggestions from proven defects; taste alone is not a finding.
  2. Establish scope. Read the target and its intent from the request, merge request description or commit messages. Consider /haechi:spec using its description to establish the review scope and criteria when useful; preserve read-only scope and report unresolved product expectations rather than deciding them yourself. Report findings with a scenario and evidence for each, without code edits. Ask only when unresolved intent would materially change the review; continue independent checks.
  3. Set the focus. Default: logic defects, excess (code that does more than needed), sensitive data and safety. If the user names a focus, use that instead.
  4. Run.
    • Diff review: use /haechi:review-diff. Use a separate worker with /haechi:verify-work through /haechi:delegate when risk or separate context makes independent review useful. File and line counts may help estimate scope but do not decide delegation.
    • Area review: use the code quality rubric alongside logic review. A base or diff is not required; inspect the target, its dependencies, consumers and relevant tests. Use review-diff's evidence and reporting rules, but skip diff-specific commands and do not invent a change history. Review directly unless independent areas benefit from separate context; then use /haechi:sweep in read-only mode with the same area scope. Read-only units produce findings, never commits.
  5. Report. Use the evidence and content requirements in /haechi:review-diff; choose the clearest presentation. State what was inspected and which checks could not be performed. "No defects found in the inspected scope" is valid; it does not certify unchecked behavior or turn missing verification into a pass.
  6. Follow-up. Fix nothing during the review. The user decides what to fix; authorized fixes follow the common work contract and relevant repair, implementation or restructuring methods. Posting merge request comments or filing issues is on the ask list.

Optional: are the tests strong enough?

Only when the user asks. Copy the repo to a temporary folder outside the repository and run /haechi:mutation-check there with --base <base> so it mutates the reviewed change, never in the working tree. Report surviving mutants as open questions or test defects, per review-diff.

Extra DONE evidence

Every defect has file:line and a scenario (input or state, then the wrong result). "No defects" is a conclusion too. The working tree is unchanged — git status shows nothing new from the review.

Delegation contracts

These compatible entries identify delegated methods, not an exclusive workflow or a mandatory agent sequence. Packet assembly also supplies the canonical common work contract. The main session fills the named context fields and task-specific acceptance criteria; detailed procedures and principle bodies stay in their referenced files. read-only means repository files remain unchanged; authorized temporary or report outputs must be named separately in artifactPaths.

{
  "review": {
    "skills": [
      "verify-work",
      "review-diff",
      "impact-map"
    ],
    "principles": [
      "evidence-levels",
      "behavior-tests"
    ],
    "checks": [
      "agent-defects",
      "sibling-paths",
      "dry",
      "yagni",
      "kiss"
    ],
    "writeScope": "read-only",
    "contextRequired": [
      "repository",
      "requirements",
      "base",
      "changedFiles",
      "priorEvidence"
    ],
    "completion": [
      "Report concrete defects with expected and observed behavior and affected consumers.",
      "Distinguish passing checks, failures, missing proof and unresolved expectations."
    ]
  }
}

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.