| name | log-triage |
| description | CI/build/runtime error log analysis with root-cause detection, prioritized fix plans, and verification checklists |
| version | 1.0.0 |
| author | ETHPanda |
| tags | ["debugging","ci","error-analysis","troubleshooting","logs","devops","testing"] |
| triggers | [{"type":"keyword","keywords":["error","failed","debug","ci failure","build failed","log analysis","troubleshoot","crash","exception"],"priority":85},{"type":"pattern","patterns":["(?i)(analyze|check|debug|fix) .*(error|failure|log|crash)","(?i)(ci|build|test) .*(failed|error|broken)","(?i)(what|why) .*(failed|error|broke)"],"priority":80},{"type":"intent","intent_category":"debugging_logs","priority":90}] |
| parameters | [{"name":"log_file","type":"string","required":false,"description":"Path to log file or raw log text"},{"name":"command","type":"string","required":false,"description":"Command that was running when error occurred"},{"name":"environment","type":"string","required":false,"description":"Environment details (OS, versions, CI provider)"}] |
| prerequisites | {"env_vars":[],"skills":[]} |
| composable | true |
| persist_state | false |
| scripts | {"enabled":true,"working_directory":"./scripts","definitions":[{"name":"log_triage","description":"Analyze error logs and generate fix plan","type":"python","file":"log_triage.py","timeout":30}]} |
Skill: Log Triage & Fix Plan (Enterprise Debugging)
Source: ETHPanda
Track: Enterprise & Team Skills → Debugging
Goal: Turn raw error logs into (1) root-cause hypotheses, (2) reproducible steps, (3) prioritized fix plan, (4) verification checklist.
When to use
Use this skill when you have:
- CI failure logs (GitHub Actions, GitLab CI, Jenkins, etc.)
- Runtime crashes, stack traces, uncaught exceptions
- Build failures (TypeScript, Python, Rust, Go, Java, Solidity/Hardhat)
- Dependency / environment / permission / network issues
Required inputs (ask user to provide)
- The error log (full text preferred)
- What command was running (e.g.
pnpm test, hardhat test, pytest, cargo test)
- Environment (OS, Node/Python/Rust versions, CI provider if any)
- Recent changes (commit, dependency bumps, config edits) if known
Output format (must follow)
- 1-paragraph summary of what failed and where
- Error signature(s) (key lines you matched)
- Root-cause candidates (ranked, with confidence and reasoning)
- Fix plan (ranked, each step includes exact commands or code areas to check)
- Verification checklist (how to confirm the fix)
- If blocked: minimal questions to unblock (max 3)
Triage procedure
- Extract the first fatal error (often earlier than the last line)
- Classify into categories:
- Syntax/type errors
- Missing dependency / wrong version
- Build toolchain mismatch
- Env var / secret missing
- Permission / path / file system
- Network / rate limit / RPC issues
- Test flakiness / timing
- Match known patterns (e.g. "Cannot find module", "TypeError", "revert", "out of gas", "403/429", etc.)
- Propose minimal reproduction locally
- Provide safest fix first (pin version / clean install / config correction) before invasive refactors
Constraints
- Don't hallucinate files that don't exist; when unsure, say "check whether X exists".
- Prefer reversible changes (lockfile pin, config tweak) over broad refactors.
- Always include a verification step (rerun exact command / CI workflow).
