Help the user systematically identify and categorize failure modes in an LLM pipeline by reading traces. Use when start…
Help the user systematically identify and categorize failure modes in an LLM pipeline by reading traces. Use when starting a new eval project, after significant pipeline changes (new features, model switches, prompt rewrites), when production metrics drop, or after incidents.
Use npx skills add to install this skill into the selected agent. Phase 0 commands are generated from source rules, not verified.
npx skills add https://github.com/hamelsmu/evals-skills/blob/main/skills/error-analysis/SKILL.md -g -a codex -yDrop `-g` to install project-locally
No fixes until the root cause is real.
No fixes until the cause is real.
Use when a security incident has been detected or declared and needs classification, triage, escalation path determinat…
| this skill · 当前error analysis | GStack Investigate | Incident Investigate | incident response | |
|---|---|---|---|---|
| Rating · 评分 | — | 4.8 | 4.8 | — |
| Stars · 星标 | 1.3k | 80k | 10k | 14k |
| Risk · 风险 | Medium risk · 中风险 | Low risk · 低风险 | Low risk · 低风险 | Medium risk · 中风险 |
| Best for · 最适合 | 深度研究 | No fixes until the root cause is real. | Root-cause analysis | Use when a security incident has been detected or declared and needs classification, triage, escalation path determinat… |
| Not for · 不适合 | Workflows that require stronger human review than this catalog entry documents. | Workflows that require stronger human review than this catalog entry documents. | Quick cosmetic fixes | Workflows that require stronger human review than this catalog entry documents. |