Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →reinamaccredy/maestro-audit
Audit a project using Maestro read-only: use for code review, architecture review, deepening opportunities, backlog proposals, harness-improvement findings, or repo-wide improvement audits without fixes.
Passed 17 of 20 test cases with the skill, and 3 of 20 without it.
and 25 more in the full report
16 fixed4 passed either way3 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 16 test cases the skill fixed. Same model, same prompt, with and without the skill.
We are running a tech debt sweep on the `auth-middleware`. We want to run a focused reach-ladder pass with the strictest session mode to propose replacing lower-rung code entirely. Which CLI command triggers this specific pass, and what mode parameter should the session use?
| Checked | Without skill | With skill |
|---|---|---|
| use the maestro lean audit command | Missed | Passed |
| specify the ultra lean mode | Missed | Passed |
The instructions your model reads when the skill loads.
Use this for repo-wide improvement audits. The audit is agent work; Maestro only stores, merges, and surfaces proposals. Architecture reviews that write an HTML report use the OS temp directory only, not the repo.
Activate with a known session id:
maestro hook record --event skill_activation --skill maestro-audit --session <session_id>
Recipe checkpoint: Maestro's main workflow is the loop. Use maestro status
for current state and maestro loop next as the read-only router when the next
recipe is not obvious. Audit work uses maestro loop show audit. Use that
recipe as the shape for perceive -> choose -> act -> observe -> learn ->
continue: read the bounded surface, choose a falsifiable probe, run read-only
checks, observe findings, record durable proposals, then return the next audit
or hard stop. Writes still use the existing Maestro verbs named by the recipe.
Custom card/run recipes are allowed only when no shipped recipe fits, and must
use the same six phases, current Maestro verbs, hard stops, and continue output.
Rule: loop next recommends; outcome/proof/memory verbs write. Use
maestro loop next --chain to explain current chain position without writing,
maestro loop outcome to preserve structured attempt outcomes and transition
receipts after native work, maestro loop trace <card> to audit card-scoped
receipts, and maestro loop improve for read-only proposals over sourced
outcomes. Audit may file explicit harness or memory proposals. Do not use silent
recipe mutation, hidden stores, hidden schedulers, or proof/QA bypass.
Do not implement, edit code, or change repo artifacts during this skill run.
Produce proposals only. Temp HTML architecture reports are allowed when the
architecture-review branch asks for them.
Audit findings are backlog-only during witness sign-off. File durable proposals
with maestro harness propose; they do not become close blockers unless they
invalidate the accepted contract, proof, QA, or risk-tier policy.
maestro status, maestro harness list --all, active
features, active tasks, decisions, and repo instructions. Pick the boundedtest cases passed, vs 3 / 20 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.