Rigor Debug / Rigor Audit skill for deep learning research work. Use when the user pastes a traceback, terminal error, CUDA OOM, checkpoint load failure, shape mismatch, NaN loss symptom, or training failure and wants conservative diagnosis before any patching, with debug fixes clearly separated from research contributions. Do not use for broad refactoring, speculative adaptation, automatic exploratory patching, or general repository familiarization.
How to install safe-debug?
lllllllama/rigorpilot-skills/safe-debug
$npx -y skills add lllllllama/rigorpilot-skills --skill safe-debugInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Run `npx skills use "https://github.com/lllllllama/rigorpilot-skills" --skill "lllllllama/rigorpilot-skills/safe-debug"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/lllllllama/rigorpilot-skills" that are relevant to the current task. Run `npx skills add "https://github.com/lllllllama/rigorpilot-skills"` and select the relevant skills, then follow their instructions.
SKILL.md
| 1 | # safe-debug |
| 2 | |
| 3 | Use this as the Rigor Debug / Rigor Audit skill. The installed slug remains |
| 4 | `safe-debug` for compatibility. |
| 5 | |
| 6 | Use the shared operating principles in |
| 7 | `../../references/agent-operating-principles.md`; this skill should guide |
| 8 | conservative diagnosis without blocking the model from finding the local root |
| 9 | cause. |
| 10 | |
| 11 | ## When to apply |
| 12 | |
| 13 | - The user provides a traceback, terminal error, or concrete training or inference failure symptom. |
| 14 | - The user wants diagnosis, root-cause narrowing, and minimal patch suggestions before code is changed. |
| 15 | - The user wants a safe debug flow with explicit human approval before mutation. |
| 16 | |
| 17 | ## When not to apply |
| 18 | |
| 19 | - When the user wants a broad repository walkthrough without an active failure. |
| 20 | - When the task is speculative experimentation or code adaptation. |
| 21 | - When the user is asking for a large refactor or readability rewrite. |
| 22 | |
| 23 | ## Clear boundaries |
| 24 | |
| 25 | - Diagnose first. |
| 26 | - Do not modify repository code by default. |
| 27 | - If a patch is needed, propose the smallest fix and require explicit approval first. |
| 28 | - Escalate savepoint or branch creation before medium-risk or high-risk changes. |
| 29 | - A debug fix is not automatically a research contribution; if it changes |
| 30 | experiment meaning or comparability, say so explicitly. |
| 31 | |
| 32 | ## Output expectations |
| 33 | |
| 34 | - `debug_outputs/DIAGNOSIS.md` |
| 35 | - `debug_outputs/PATCH_PLAN.md` |
| 36 | - `debug_outputs/status.json` |
| 37 | |
| 38 | ## Notes |
| 39 | |
| 40 | Use `references/debug-policy.md`, `../../references/research-rigor-principles.md`, and the shared `../../references/research-pitfall-checklist.md`. |
Related
6 picksType
- azure-diagnosticsDebug Azure production issues on Azure using AppLens, Azure Monitor, resource health, and safe triage.Jul 2026
- appinsights-instrumentationGuidance for instrumenting webapps with Azure Application Insights. Provides telemetry patterns, SDK setup, and configuration references.Jul 2026
- diagnosing-bugsDiagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.Jul 2026
- systematic-debuggingUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesJul 2026
- golang-error-handlingIdiomatic Golang error handling — creation, wrapping with %w, errors.Is/As, errors.Join, custom error types, sentinel errors, panic/recover, the single…Jul 2026
- golang-performanceGolang performance optimization patterns and methodology - if X bottleneck, then apply Y.Jul 2026
