agentleFS
Sign inStart free
davidscotttuftsdavidscotttufts/hyvemind/skills/review-change/SKILL.md

review-change

Skill0 starsChanged 20 days agoApache-2.0View on GitHub

Senior code review of the current branch: runs the full hyvemind gate first, scans the diff for the project's banned patterns, then reviews for correctness, security/authorization, data integrity, error honesty, tests and simplicity. Findings are severity-ordered with concrete failure scenarios; ends with a GO / NO-GO for opening a PR.

29 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use plan-change, implement-change and 6 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.
  • BanKnightSkill

    plan-change

    将已完成 spec/design 的 change 拆成可执行 plan 与 tasks。用户要把设计落成执行计划、任务清单或实现步骤时使用。

    Also in 7 projects that use this

  • Codagent-AISkill

    implement-change

    Implements an approved change in one continuous session, with tiered test runs. Use when asked to implement a change from its approved definition, or to invoke codagent:implement-change.

    Also in 7 projects that use this

  • MaheshAwasareSkill

    write-runbook

    Generate an operational runbook for a service from its code, config, and dependencies — what to do when it breaks, where to look, how to roll back, who to escalate to. Use proactively before a service goes to production, after a major incident, or when handing off operational ownership.

    Also in 3 projects that use this

  • testified-ossSkill

    incident-triage

    Triage production incidents with severity assessment, root cause hypothesis, impact analysis, and escalation recommendations. Use when analyzing outages, errors, or production issues.

    Also in 4 projects that use this

  • srclightSkill

    refactor-safely

    This skill should be used when the user asks to "refactor", "rename", "move", "change a function signature", "extract a method", "delete unused code", "clean up", or any modification that could break callers or dependents. Provides a pre-change safety checklist using srclight's graph tools to assess blast radius before touching code.

    Also in 3 projects that use this

  • ShopifySkill

    prepare-release

    Prepare the changeset-release/main branch for release by syncing version.ts files and running pnpm install

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 25 edited versions we read keep these.

  • 5

    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:

    The author never grades their own change: If you or an agent you directed wrote it, hand the review to a fresh agent or session with only …

    line 8
  2. Very specific:

    Run {{cli}} gate (or the project's external gate).

    line 13
  3. Distinctive:

    Read the whole change, including the tests.

    line 27
  4. Distinctive:

    Don't report style preferences a formatter or linter would catch.

    line 39
  5. Distinctive:

    Before reporting, try to disprove each finding (is the path reachable?

    line 43
  6. Distinctive:

    Every finding needs a scenario you could turn into a failing test.

    line 61

“Common” is measured: kept by at least 6 in 10 of the 25 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 change

The author never grades their own change. If you or an agent you directed wrote it, hand the review to a fresh agent or session with only the spec and the diff.

1. Gate first

Run {{cli}} gate (or the project's external gate). If it is not green, that is the first and blocking finding. Report the failing gate's output and stop. Don't review code that won't merge.

2. Mechanical scan of the diff

Get the diff against the base branch: git diff $(git merge-base HEAD origin/main), or whatever base the project uses. Check every added line against review.bannedPatterns in hyvemind.json (each has a pattern and a message). Each hit is a BLOCKER unless the line carries a justification the user has already accepted.

3. Read the diff for defects

Read the whole change, including the tests. Look for:

AreaAsk
CorrectnessOff-by-one, null/empty, concurrency, ordering, time zones, retries that run twice
Error honestyDoes any failure come back as success (empty list, default value, swallowed catch)? Invalid input should be a typed error.
SecurityIs every new read/write authorized by the same check as its neighbours? Injection, path traversal, SSRF, secrets in logs, unsafe deserialization.
Data integrityOrphans on delete, non-idempotent writes, missing transactions, migrations without rollback
TestsDoes each new test fail if the behaviour breaks? If in doubt, probe it ({{skill:test-necessity}}). Tests should assert behaviour, not incidental shape.
SimplicityCode that duplicates an existing helper, dead code, speculative abstraction
ScopeChanges unrelated to the stated goal belong in their own PR

Don't report style preferences a formatter or linter would catch.

4. Refute pass

Before reporting, try to disprove each finding (is the path reachable? does a guard elsewhere already handle it?) and each "looks fine" (what input would break it?). Drop findings that don't survive; reopen areas whose "fine" doesn't.

5. Report

VERDICT: GO | NO-GO
Gate: green · tree <hash> (from `{{cli}} evidence`)
Coverage: <audited> audited + <out-of-scope> out of scope + <generated/dead> = <total> files in the diff

[BLOCKER] <file>:<line> — <one-line defect>
  Scenario: <inputs/state> → <wrong outcome>
  Fix: <specific change>
[MAJOR] …
[MINOR] …

Every finding needs a scenario you could turn into a failing test. If you can't write one, drop the finding. If the coverage sum doesn't equal the total, a file was silently skipped: review it.

⛔ Any BLOCKER means NO-GO: go back to implementation, then review again.

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