maestro-audit — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited maestro-audit (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Use this for repo-wide improvement audits. The audit is agent work; Maestro only stores, merges, and surfaces proposals.
Activate with a known session id: maestro hook record --event skill_activation --skill maestro-audit --session <session_id>
Do not implement, edit code, or change repo artifacts during this skill run. Produce proposals only.
maestro status, maestro harness list --all, activefeatures, active tasks, decisions, and repo instructions.
tests, scripts, and shipped embedded resources relevant to the finding. Sweep every lens so coverage is checkable, not just whatever surfaced first: correctness, security, performance, test coverage, tech debt, dependencies, developer experience, docs. The tech-debt lens includes the reach-ladder (HARNESS Code style): code a lower rung -- stdlib, native platform, an installed dependency, a one-liner -- already covers. The session lean mode tunes how strictly to propose these (maestro lean): ultra proposes replacing such code, full/lite propose the cheaper form, off skips the reach-ladder lens. maestro lean audit runs the focused, mode-adjusted reach-ladder pass; this skill still only proposes (no edits, no markers).
(re-read the code, re-run the command). Drop findings that do not survive.
accepted, dismissed, measured, or covered by active tasks.
maestro harness propose(signatures: reference/cli.md). Use one stable --topic per finding so the verb merges repeats, and end the --evidence text with a leverage estimate: impact/effort/confidence: <H|M|L>/<H|M|L>/<H|M|L>.
Each proposal needs concrete evidence: file paths, line numbers, command output, or exact artifact names, plus the closing impact/effort/confidence estimate (H, M, or L each) so the backlog ranks without re-deriving it. Do not file style opinions without a repo-specific impact and a way to verify the improvement.
Pipeline: [maestro-audit] -> maestro harness apply -> maestro-card
Next: proposals filed -> inspect with maestro harness list; accepted proposals spawn normal tasks through maestro harness apply <id>.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.