agentleFS
jellydnjellydn/my-ai-tools/skills/code-quality-review/SKILL.md

code-quality-review

Skill127 starsChanged 26 days agoMITView on GitHub

Audit a diff for structural maintainability and unnecessary complexity.

23 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use dependency-cve-sweep, fleet-standard-adoption and 5 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.
  • berntpoppSkill

    dependency-cve-sweep

    Use when handling Dependabot PRs, CodeQL/Trivy alerts, or doing a routine dependency/CVE sweep for this repo.

    Also in 12 projects that use this

  • berntpoppSkill

    fleet-standard-adoption

    Use when bringing this repo into compliance with a GeneFoundry fleet standard (tool-naming, response-envelope, container-hardening, versioning, MCP transport, or logging) or closing its tracking issue.

    Also in 12 projects that use this

  • IgorGanapolskySkill

    ci-failure-triage

    Triage a CI / PR check failure by READING the failure body before forming any hypothesis. Triggered whenever a required check is red, a PR is BLOCKED, a merge won't land, or you're about to call a failure "transient", "flaky", "stale", or "orphaned". Prevents dismissing a real failure (e.g. real CodeQL security alerts) as noise.

    Also in 12 projects that use this

  • Microsoft AzureSkill

    release-readiness

    Complete ODIN release documentation and validation checks before declaring user-visible work complete, preparing a release, or opening the Release-to-main pull request.

    Also in 12 projects that use this

  • SentrySkill

    security-review

    Finds exploitable application security vulnerabilities in code changes. Use for Warden security scans, appsec review, OWASP-style checks, authentication or authorization bugs, injection, XSS, SSRF, path traversal, secrets, unsafe crypto, webhook verification, open redirects, or sensitive data exposure.

    Also in 13 projects that use this

  • +2 more from berntpopp

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 18 edited versions we read keep these.

  • 7

    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:

    Treat repository conventions and existing canonical helpers as the source of truth.

    line 20
  2. Distinctive:

    Prefer a small set of high-confidence structural findings over cosmetic notes.

    line 22
  3. Distinctive:

    Do not propose a large refactor unless its benefit is clear and it stays within the change's ownership boundary.

    line 23
  4. Distinctive:

    Check each meaningful change for:

    line 27
  5. Distinctive:

    Use existing complexity and duplication reports as evidence when available, not universal pass/fail thresholds.

    line 53
  6. Distinctive:

    Treat these as blockers unless the implementation has a clear justification:

    line 66
  7. Distinctive:

    Do not block on personal style, hypothetical future needs, or a rewrite that is only differently complex.

    line 75

“Common” is measured: kept by at least 6 in 10 of the 18 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

Code Quality Review

Review the current branch diff for structural quality. Preserve behavior, but actively look for a simpler design that deletes concepts, branches, wrappers, or layers instead of moving complexity around.

Review Boundaries

  • Review only the changed code and the surrounding code needed to judge it.
  • Treat repository conventions and existing canonical helpers as the source of truth.
  • Skip formatter and linter findings that automation already reports.
  • Prefer a small set of high-confidence structural findings over cosmetic notes.
  • Do not propose a large refactor unless its benefit is clear and it stays within the change's ownership boundary.

Review Criteria

Check each meaningful change for:

  1. Simpler structure — Can a different model or ownership boundary remove branches, modes, helpers, or layers?
  2. Cohesive control flow — Flag scattered special cases, repeated conditions, deep nesting, and mixed responsibilities.
  3. Useful abstractions — Flag thin wrappers, generic magic, speculative extension points, and duplicate helpers.
  4. Clean boundaries — Keep feature logic in its canonical module and make types and invariants explicit.
  5. Reasonable file size — Treat a change that pushes a file above 1,000 lines as a decomposition signal unless the file has a clear structural reason to stay whole.
  6. Sound orchestration — Flag unnecessarily sequential independent work and updates that can leave related state partially applied when a clearer atomic design exists.
  7. Legibility — Prefer direct, boring code with clear names and comments that explain only non-obvious reasons.

Boundary and state checks

Apply these checks only where the diff touches the relevant contract:

  • Input and domain model — Validate external data at entry and carry that validated contract internally. Prefer existing types that exclude invalid combinations over repeated guards or casts that hide a broken model. Do not remove checks at a separate trust boundary or require a type-system redesign for clear, local code.
  • Retries and partial failure — For operations that can be retried, check the outcome after a second run and after interruption between side effects. Flag duplicate effects or incomplete recovery against the operation's contract; do not assume that every operation must be idempotent.
  • Mutable ownership — Before proposing locks or serialized execution, check whether independent actors can own separate state. When state must be shared, identify the writer and the mechanism that enforces safe updates. Preserve a correct existing transaction or synchronization model rather than forcing a single-writer rewrite.

Use existing complexity and duplication reports as evidence when available, not universal pass/fail thresholds. Recommend a split or shared helper only when it improves cohesion or gives one business decision one owner; similar syntax alone is not enough. Do not require Crapper, Dryer, or new tooling to complete a review.

Finding Bar

Report a finding only when you can name:

  • the file and relevant hunk;
  • the concrete maintenance cost or failure mode;
  • the smallest practical remedy; and
  • why the remedy is better than the current design.

Treat these as blockers unless the implementation has a clear justification:

  • a structural regression or avoidable boundary leak;
  • a file newly crossing 1,000 lines without useful decomposition;
  • ad hoc branching added to an already busy flow;
  • duplicated logic where a canonical helper exists;
  • an abstraction, cast, or optional contract that adds indirection without clarity; or
  • a clear simplification that removes substantial incidental complexity.

Do not block on personal style, hypothetical future needs, or a rewrite that is only differently complex.

Output

Order findings by impact:

  1. Structural regressions and simpler designs
  2. Control-flow, boundary, abstraction, and type problems
  3. File-size, modularity, and legibility concerns

For each finding, use:

[severity] file:line — finding
Impact: concrete cost or risk
Fix: smallest practical remedy

End with APPROVE when no blocking finding remains, or CHANGES REQUESTED with the blocker count. Be direct and respectful.

Attribution

Boundary and state checks and metric limits are adapted from Robert C. Martin's ub-mode — Architecture, Crapper, and Dryer.

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.