Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ReinaMacCredy/maestro --skill maestro-diagnosegit clone --depth 1 https://github.com/ReinaMacCredy/maestroWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/reinamaccredy/maestro/maestro-diagnose)<a href="https://agentmods.dev/skills/reinamaccredy/maestro/maestro-diagnose"><img src="https://agentmods.dev/badge/skills/reinamaccredy/maestro/maestro-diagnose/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/reinamaccredy/maestro/maestro-diagnose"><img src="https://agentmods.dev/badge/skills/reinamaccredy/maestro/maestro-diagnose.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.00518 |
| Opus 5 | $0.00023 | $0.00259 |
| Sonnet 5 | $0.00009 | $0.00104 |
| Haiku 4.5 | $0.00005 | $0.00052 |
Grade A, and why
maestro-diagnose scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
maestro-diagnose
Find the cause; change nothing. Any tier may use it.
- Reproduce. Capture the exact failing command or input and its output. No reproduction: say so and stop; do not guess from symptoms.
- Localize. Bisect the surface: which layer, module, commit, or input
feature flips the behavior. Prefer evidence (logs, minimal harnesses,
git bisect) over reading and speculating. - Reduce. Shrink to the smallest input or state that still fails.
- Deliver. Root cause plus its evidence chain, the reduced reproduction,
and the smallest fix direction. With a work item:
maestro work note <id>carries the one-line cause and the reproduction command; in a bundle, also the NOTES.md Current State. Otherwise the conversation.
Discipline
- Every probe follows a written hypothesis, "I believe X causes this because Y", then the minimal test that can falsify it. No hypothesis, no probe.
- Recall first:
maestro search "<symptom>"andmaestro attention. A priorfailed:note on the same mechanism is evidence about which hypotheses are already refuted. - Three refuted hypotheses in a row: stop treating this as a local bug.
Question the architecture; a structural cause or an unsettled design
decision is
maestro-designwork, not a fourth probe. - Red flags that mean return to step 1: proposing a fix before the cause is known, changing several things at once, "probably X" without a check.
Diagnosis is read-only toward production code: instrumentation and harnesses live outside production paths or are reverted before delivery.
An explicit fix request routes by target and tier (maestro-bundle tier
rule): a quickfix or Light fix proceeds directly with the diagnosis as its
first step; Full-tier scope is maestro-work inside the bundle, with the
diagnosis as its first phase; a prototype fix belongs to maestro-explore.
Never downgrade a requested fix to analysis.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 44 lines · 46 tokens per session scan A 39d07a6be523
maestro-diagnose is a skill published in the GitHub repository ReinaMacCredy/maestro (232 stars, last pushed 6d ago), licensed MIT. It adds 46 tokens to every session and 518 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-05.
Other skills, from other repositories
pn-session-retro
Run a blameless retrospective on an agent transcript. Classify mistakes under a 5-code taxonomy, emit a structured report at docs/refs/retros/, and recommend concrete hardening. Use after a session that felt off, before a quarterly audit, or via the /pn-retro command. v1 is manual-only; cross-session pattern counters…
pn-api-probe
Fetch live API, version, and enum facts from the target runtime or stack before planning. Use in Phase 0 of pn-design and pn-build when the stack may have changed since training data.
pn-systematic-debugging
Root cause analysis with triage mode — Phase 0 feedback loop, investigate first, one question max, then isolate, hypothesize, confirm. Outputs TDD fix plan with RED-GREEN cycles; optional GitHub issue via GitHub MCP. Use when debugging a failure or bug.
pn-discipline-philosophy
Defines engineering discipline: test-first, root-cause before fix, minimal change, measure-before-optimize, review/second look, evidence over guess. Use when implementing features, debugging, planning, or establishing development practices. Aligns with TDD, systematic debugging, and RCA practices (current).
migrate-oxlint
Guide for migrating a project from ESLint to Oxlint. Use when asked to migrate, convert, or switch a JavaScript/TypeScript project's linter from ESLint to Oxlint.
adversarial-reviewer
Adversarial code review that assumes bugs exist and hunts for them. Use when asked to review code, find bugs, audit for correctness, stress-test a PR, or when someone says "tear this apart" or "what's wrong with this". Give no benefit of the doubt — every line is guilty until proven innocent.