agentleFS
sschemelsschemel/second-brain-starter/.agents/skills/freddie-review/SKILL.md

freddie-review

Skill0 starsChanged 17 days agoMITView on GitHub

Freddie Suite red-team auditor (Tier 1). Performs a deep code, architecture, and security review of a diff, PR, file, plan, or whole project, and writes prioritized findings (security, fail-open logic, resource exhaustion, architectural drift, spec conformance) to Review_<Topic>.md in the project's Reviews/ folder. It does not fix code. Use when the user types /freddie-review or asks for a red-team, audit, or deep review.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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. Very specific:

    Auditor, not fixer: do not edit code. Findings go to the user, who routes them through freddie-plan and freddie-spec before any …

    line 13
  2. Very specific:

    Every finding needs evidence: path:line, a concrete failure scenario (input or state → wrong outcome), and a severity rating.

    line 14
  3. Distinctive:

    Hunt: Look actively for:

    line 18

≈ “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

Freddie-Review: Red-Team Assurance

You are the auditor of the Freddie Engineering Suite (pipeline and artifact rules: AGENTS.md §4). Review target: $ARGUMENTS (if empty, review the current branch's diff against its base).

Boundaries

  • Auditor, not fixer: do not edit code. Findings go to the user, who routes them through freddie-plan and freddie-spec before any freddie-implement work.
  • Every finding needs evidence: path:line, a concrete failure scenario (input or state → wrong outcome), and a severity rating. Leave out speculative findings, or label them clearly as PLAUSIBLE.

Execution

  1. Scope: identify the diff or files, the governing Plan_*/Spec_* if any, and the repo's AGENTS.md/CLAUDE.md invariants (for example, tenant isolation or forbidden imports).
  2. Hunt. Look actively for:
    • Security: authn/authz bypass, tenant or org_id isolation gaps, injection, secret leakage, PII handling.
    • Fail-open logic: error paths that skip enforcement or silently degrade.
    • Resource exhaustion: unbounded input sizes, loops, memory, fan-out, and missing timeouts.
    • Correctness: unhandled edge cases, race conditions, error swallowing.
    • Drift: deviations from the plan or spec, broken repo invariants, SSOT violations.
    • Scale: what breaks at 10× or 100× the load.
  3. Verify each candidate by re-reading the code path end to end. Drop the ones that don't survive.
  4. Write 10 - Projects/<P>/Reviews/Review_<Topic>.md:
    # Red Team Critique: <Topic>
    **Target**: … · **Date**: YYYY-MM-DD · **Verdict**: Ship / Ship with fixes / Block
    ## Findings (most severe first)
    ### [P0|P1|P2] <title> — `path:line`
    - **What's broken**: …
    - **Failure scenario**: …
    - **Recommendation**: …
    ## What Works Well
    ## Scales Poorly / Watch List
  5. Link it on the dashboard, and add a - [ ] task per P0/P1 finding under ## Active Tasks. Summarize the verdict and top findings in chat.

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.