debug-trace
/debug-trace
Zapisuje dowody, odrzucone hipotezy i kolejne kroki długiego debugowania w katalogu .debug/traces.
- Autor
- yuri-xyz ↗
- Repozytorium
- skills ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 03.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 2
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 03.11.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Zapisuje dowody, odrzucone hipotezy i kolejne kroki długiego debugowania w katalogu .debug/traces.
Oryginalny opis (EN)
Use when a long or multi-session debugging effort needs its evidence, ruled-out theories, and next steps recorded under `.debug/traces`.
Jak działa
- When To Use
- Workflow
- Trace Format
- Boundaries
Podgląd SKILL.md
--- name: debug-trace description: Use when a long or multi-session debugging effort needs its evidence, ruled-out theories, and next steps recorded under `.debug/traces`. --- # Debug Trace Use this skill to keep a durable investigation note for the current bug. The trace should make the next agent run start from what is already known instead of rediscovering the same facts. ## When To Use Use this skill automatically during bug investigations when new information materially changes the debugging state, including: - A log, test, reproduction, or command output provides key evidence. - Code inspection reveals a likely boundary, branch, invariant, or state transition involved in the issue. - A hypothesis is confirmed, weakened, ruled out, or superseded. - The next best debugging direction changes. Do not wait for an explicit `/debug-trace` request when those events happen. ## Workflow 1. Identify the current problem being investigated from the conversation, repo context, logs, failing command, or active debug instrumentation. 2. Create `.debug/traces/` if needed. 3. Reuse an existing trace file when it clearly matches the same problem. Otherwise create a new Markdown file with a short kebab-case name, such as `.debug/traces/niri-resume-hang.md`. 4. Read the existing trace before editing it, and preserve useful prior evidence. 5. Add or update only concrete investigation state: - Symptom and scope. - Evidence observed from logs, commands, tests, screenshots, reproduction steps, or code inspection. - Things narrowed down, including paths, branches, state transitions, inputs, outputs, and timing. - Theories ruled out or deprioritized, with the evidence that made them less likely. - Best next directions to keep narrowing the problem. 6. Keep speculative notes clearly labeled as hypotheses. Do not present guesses as evidence. 7. Avoid secrets, tokens, private keys, credentials, cookies, personal data, and large payload dumps. Redact or summarize sensitive values. 8. Report the trace file path and the main update made. ## Trace Format Prefer this structure unless the existing file already has a useful shape: ```markdown # Short Problem Name ## Problem - Current symptom: - Reproduction or trigger: - Scope: ## Evidence - ... ## Narrowed Down - ... ## Ruled Out - [ ] Theory: Evidence: ## Next Directions - ... ``` Use checkboxes in `Ruled Out` only when they help scan open vs closed theories. It is fine to write normal bullets when that reads better. ## Boundaries - Trace files are investigation notes, not source-of-truth docs or permanent project plans. - Keep trace files under `.debug/traces/*.md`. - Do not edit generated output or runtime state just to update a trace. - Do not delete debug logs, markers, or trace files as part of this skill. Cleanup belongs to `/debug-cleanup`.
Pliki w paczce (2)
- SKILL.md 3 KB
- agents/openai.yaml 1 KB
Wersje
W bazie nie ma innych skilli o nazwie debug-trace.
Migawka bazy z 13.11.2026 · wersja paczki 4d1a3ae8bb09 (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100investigation-state
Zapisuje kontekst sesji debugowania (hipotezy, dowody, odrzucone podejścia), by można go było wznowić lub przekazać w kolejnych sesjach. - 100linear-debug-bundle
Dostarcza zestaw narzędzi do debugowania integracji z Linear: logowanie, śledzenie wywołań API i własne narzędzia diagnostyczne. - 100trace
Sprawdza równolegle konkurujące hipotezy przyczyny i zbiera dowody, by ustalić, dlaczego coś się wydarzyło. - 100playwright-evidence
Zbiera dowody z testów E2E (zrzuty ekranu, wideo, trace'y) i zapisuje je w katalogu .copilot-tracking dla agentów, z konfiguracją Playwright, przechwytywaniem przy błędach i manifestem dowodów. - 100debug
Debuguje systematycznie błędy w integracji Home Assistant: bada wyjątki, analizuje stack trace i naprawia usterki. - 100stack-trace-analysis-symbolication-source-maps
Odszyfrowuje zaciemnione ślady stosu za pomocą symboli debugowania, raportów awarii i map źródłowych.