issue-investigator
Investigate software issues or incidents. Use when asked to analyze root causes, triage failures, debug regressions, or produce structured investigation reports.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 2
- Forks
- 0
01 前言 Front Matter
- name
- issue-investigator
- description
- Investigate software issues or incidents. Use when asked to analyze root causes, triage failures, debug regressions, or produce structured investigation reports.
02 资源文件 1
- 参考文档agents/openai.yaml189 B
03 SKILL.md 正文
Issue Investig8tor
Lead a structured investigation of a user-provided issue. Use deep internal reasoning but do not reveal chain-of-thought. Provide concise, actionable outputs.
Core Rules
- Do not reveal chain-of-thought or internal reasoning. Provide conclusions and evidence only.
- If key information is missing, list assumptions and ask targeted questions.
- Keep hypotheses distinct and non-overlapping.
Workflow
- Triage and clarify
- Extract: symptoms, expected vs actual behavior, environment, repro steps, logs/errors, timeline, and recent changes.
- Identify missing context and ask for it if needed.
- Generate hypotheses
- Produce up to 8 plausible, distinct root-cause hypotheses. Scale count to issue complexity and evidence available.
- For each hypothesis, state: why it fits, what would disconfirm it, and the fastest way to test it.
- Parallel investigation plan
- Spawn up to 8 sub-agents, one per active hypothesis. Scale sub-agent count to complexity; use fewer for narrow issues.
- Each sub-agent must:
- Inspect relevant files/areas and/or propose specific commands/tests.
- Collect evidence for/against the hypothesis.
- Return a short report with: evidence, confidence (0-100), and next action.
- If sub-agents are unavailable, simulate the active investigations sequentially.
- Synthesis
- Compare evidence across reports.
- Eliminate weak hypotheses and identify the most likely cause(s).
- Provide a recommended fix and a minimal verification plan.
Output Format
- Summary: 2-4 sentences.
- Hypotheses (ranked): H1-H8 max with 1-2 lines each and confidence.
- Evidence highlights: bullets keyed to hypotheses.
- Most likely cause(s): 1-2 bullets.
- Recommended fix: 1-3 bullets.
- Verification plan: 2-4 bullet steps.
- Questions / missing info (if any).
04 同仓库更多 Skills
writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code
workflow-from-chats
Extract preferences from agent chats → skills, rules, workflow docs. For learning preferences, mining feedback, personalizing workflows.
winui3-csharp-app
WinUI 3 / Windows App SDK desktop apps in C#/XAML. MVVM, bindings, navigation, Mica/Acrylic, windowing, packaging, performance.
why
Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds. Discovers available MCPs and queries each evidence category (source control, issue tracker, long-form docs, real-time chat, infrastructure observability, error tracking, product analytics warehouse) in parallel, then returns a cited read on decisions and tradeoffs. Use how for runtime behavior.
what-did-i-get-done
Summarize authored commits over a user-specified time period into a concise status update.
weekly-review
Produce a short weekly synthesis of authored commits with highlights by bugfix, tech debt, and net-new work.