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 tmusser/ai-engineering-skills --skill diagnose-loopgit clone --depth 1 https://github.com/tmusser/ai-engineering-skillsWrote 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/tmusser/ai-engineering-skills/diagnose-loop)<a href="https://agentmods.dev/skills/tmusser/ai-engineering-skills/diagnose-loop"><img src="https://agentmods.dev/badge/skills/tmusser/ai-engineering-skills/diagnose-loop/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/tmusser/ai-engineering-skills/diagnose-loop"><img src="https://agentmods.dev/badge/skills/tmusser/ai-engineering-skills/diagnose-loop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00017 | $0.00589 |
| Opus 5.5 | $0.00007 | $0.00236 |
| Sonnet 5.5 | $0.00003 | $0.00118 |
| Haiku 4.5 | $0.00002 | $0.00059 |
Grade A, and why
diagnose-loop 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diagnose Loop
Purpose
Debug failures systematically.
When to use
Use when a command fails, behavior is wrong, a test regresses, or output does not match the spec.
Inputs
- Failure report
- Exact command or reproduction case
- Recent changes
- Logs, traces, or output
Workflow
- Freeze feature work.
- Reproduce the failure.
- Identify the exact failing command or case.
- List hypotheses.
- Instrument minimally.
- Name one active hypothesis and the observation that would falsify it; apply the smallest fix that tests it.
- Add a regression check if possible.
- Rerun the original failing command.
- Rerun relevant verification. An improved signal is not completion; the original failure and compatibility checks must pass.
- Before a second repair attempt, count the initial attempt and define the retry contract: fixed input/evaluator identities, acceptance threshold, total attempt limit, consecutive no-progress limit, revert rule, ledger location, and review trigger. If incomplete, stop with
REVIEW_REQUIRED; the initial bounded diagnosis does not require a separate loop artifact. - Record each started attempt, including failed commands and interrupted attempts: hypothesis, artifact state, evidence, outcome, and consumed budget. Change the next hypothesis only when the evidence supports it. Unchanged or worse results and tool errors consume the no-progress budget; improvement does not reset the total budget.
- Stop at acceptance, either budget limit, an unresolved interruption, or a scope/permission/evaluator change. Preserve the best known state and the failed attempt's evidence; revert only edits owned by this attempt, never unrelated user work. Reconcile uncertain external effects before retrying.
- On resume, carry forward the same loop identity, ledger, input/evaluator identities, consumed budget, and stop reason. A fresh session or subtask does not grant new attempts. Renew a spent budget or change the evaluator only through an explicit contract revision; preserve the prior ledger.
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.
- 1mo ago Changed · +8 lines · +9 tokens per session c04a70b54a57
- 1mo ago First seen · 53 lines · 8 tokens per session scan A b25c65d3200d
diagnose-loop is a skill published in the GitHub repository tmusser/ai-engineering-skills (6 stars, last pushed 5d ago), licensed MIT. It adds 17 tokens to every session and 589 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
bug-diagnosis
Use this skill to systematically diagnose the root cause of a bug before attempting a fix. Activates when a bug is reported, a test is failing, or unexpected behavior is observed. Prevents guess-and-check debugging that introduces additional bugs while chasing symptoms.
doca-flow
DOCA Flow packet steering on NVIDIA NICs/DPUs, linked into a C application: match/action pipes, ports, representors, forwarding, hairpin, counters. Trigger on packet steering, classifier, rule-matching, representor, hairpin or 5-tuple-to-queue questions even when "DOCA Flow" is not named. Also covers pipe validation…
compute-sanitizer-diagnose-cuda-bugs
Detect and diagnose CUDA correctness bugs with NVIDIA Compute Sanitizer. Routes GPU memory errors, shared-memory data races, uninitialized-memory reads, and synchronization/barrier bugs to the right sub-tool (memcheck, racecheck, initcheck, synccheck) and explains how to build, run, and read the report. Use when…
doca-debug
Use this skill when the user is debugging any DOCA symptom — a build that won't compile, a link step that can't resolve a doca symbol, a runtime call returning DOCAERROR, a silent service or tool, or a stack trace / valgrind / core dump — and needs the layered ladder (install → version → build → link → runtime →…
mcore-create-issue
Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.
doca-dpa-hl-tracer
Use this skill when the user runs docadpahltracer to capture/decode DPA-side traces at the programming-events layer (kernel entry/exit, sync points, comm primitive calls, RDMA WR submission, completion drain) — picking TRACE vs CRIT, tuning the JSON config (file-size limits + filesizelimitpolicy, thread…