agentleFS
sssemilsssemil/SKILLS/skills/brutal/brutal-project-review/SKILL.md

brutal-project-review

Skill2 starsChanged 4 months agoMITView on GitHub

Systematically review a project subsystem-by-subsystem with resumable state tracking. Creates CRITICAL/MAJOR finding tasks through the work store resolved from BRUTAL.md.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read repo rules from AGENTS.md, CLAUDE.md, TARGET.md, and referenced workflow docs.

    line 16
  2. Very specific:

    If BRUTAL.md is missing or incomplete, follow the integration setup flow.

    line 18
  3. Distinctive:

    write temporary context under the resolved review-state directory

    line 40
  4. Distinctive:

    write the subsystem report

    line 44
  5. Distinctive:

    delete temporary context before stopping or reporting

    line 45
  6. Very specific:

    Only create persisted work for validated CRITICAL and MAJOR findings by default.

    line 46
  7. Distinctive:

    Before creating a finding, search existing state and work-store artifacts for the same fingerprint and these legacy source markers:

    line 57
  8. Very specific:

    Create/update type:review-finding artifacts through the work-store adapter in its logical todo state, add progress to the parent when one …

    line 65

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Brutal Project Review

Review the project one subsystem at a time and persist actionable CRITICAL/MAJOR findings through the configured work-store adapter.

Required Context

  1. Read ../brutal-shared/integration-resolver.md.
  2. Resolve the work store and load its support module before initializing state or creating findings.
  3. Read repo rules from AGENTS.md, CLAUDE.md, TARGET.md, and referenced workflow docs.
  4. If BRUTAL.md is missing or incomplete, follow the integration setup flow.

State Rules

  • The local adapter uses its configured review-state path and local tasks.
  • Remote adapters use .brutal-workspace/review-state/ only for resumable local review state. Exclude .brutal-workspace/ through .git/info/exclude before writing it and never create local work-store artifacts.
  • Preserve legacy manifest fields when resuming older runs, but write new entries with adapter-neutral keys where practical: work_store, work_store_project, parent_ref, findings_created.

Workflow

  1. Resolve or initialize the review manifest.
  2. Discover subsystems dynamically from behavior-defining files. Exclude generated/vendor/build folders. Normalize ids as 01, 02, etc.
  3. Show the subsystem table and ask what to review: id/name/path, all, pending, restart, or cancel.
  4. For each selected subsystem:
    • mark it in_progress
    • gather relevant code, tests, migrations, APIs, configs, SQL, TOML, and docs
    • write temporary context under the resolved review-state directory
    • launch five reviewers when available: core correctness, reliability/tests, maintainability, performance/security, and simplification
    • synthesize, dedupe, validate, and severity-check findings
    • write the subsystem report
    • delete temporary context before stopping or reporting
  5. Only create persisted work for validated CRITICAL and MAJOR findings by default.

Finding Persistence

Use deterministic fingerprints:

<subsystem-id>|<canonical-file>|<line-or-symbol>|<severity>|<normalized-title>

Before creating a finding, search existing state and work-store artifacts for the same fingerprint and these legacy source markers:

Source: brutal-project-review
Legacy sources: brutal-project-review, linear-brutal-project-review, gitlear-brutal-project-review

Create/update type:review-finding artifacts through the work-store adapter in its logical todo state, add progress to the parent when one exists, and append stable refs to the review manifest. Follow provider-specific read fallbacks and prepare-before-create rules from the support module.

Finding bodies must include severity, subsystem, file/line, fingerprint, confidence, issue explanation, suggested fix, acceptance criteria, implementation notes, dependencies, and verification commands.

Final Response

Lead with findings by severity. Include work store, created/updated finding refs, manifest path, parent ref, reviewed subsystem count, remaining subsystem count, and recommend $brutal-swarm for parallel fixes or $brutal-worker for one exact finding.

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.