interrogate
Adversarially review a diff for correctness, security, regressions, and missing coverage. Use for `interrogate`, multi-agent review, challenge-this, or requests to find blind spots.
- License 许可
- MIT
- 最近更新
- 2026-10-06
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- interrogate
- description
- Adversarially review a diff for correctness, security, regressions, and missing coverage. Use for `interrogate`, multi-agent review, challenge-this, or requests to find blind spots.
02 资源文件 4
- 参考文档references/code-quality-review.md5.1 KB
- 参考文档references/lead-judgment.md3.4 KB
- 参考文档references/reviewer-prompt.md2.4 KB
- 参考文档references/rubric.md4.9 KB
03 SKILL.md 正文
Interrogate
The deliverable is a synthesized review verdict. Do not auto-apply changes.
Step 1. Determine scope
Use the caller's files or diff. Otherwise inspect the current working diff against its base branch and gather only the context needed to review it.
Step 2. State intent
Derive the intended behavior from the user's request, any PR description, and the code. State the intended behavior in one paragraph. If it is genuinely ambiguous, identify the ambiguity and its impact.
Step 3. Review independently
Read references/reviewer-prompt.md, references/rubric.md, and references/code-quality-review.md. Give the same intent, patch, context, and rubric to independent read-only agents when the host supports delegation. Ask reviewers to report only actionable findings with file and line references, trigger conditions, impact, and evidence. Use a configured comment-focused reviewer only when one is available.
When several configured agent profiles use distinct models, use them for model diversity. Otherwise treat the results as independent passes from the same model. If delegation is unavailable, do separate review passes yourself and disclose the limitation.
Do not ask reviewers to change files. Do not widen the review beyond the supplied change and its necessary context.
Step 4. Synthesize
Deduplicate findings, distinguish consensus from single-reviewer findings, and inspect every cited path yourself. Categorize each finding as Act on, Consider, Noted, or Dismissed. Give a one-line evidence-based reason for the disposition.
Output
Return Intent, Reviewers, Act On, Consider, Noted, Dismissed, and Agreement Map. Each finding includes exact location, impact, trigger, and disposition. If there are no findings, say what scope you reviewed and what the review could not establish.
04 同仓库更多 Skills
why
Trace the evidence behind an architectural choice, regression, postmortem, or measured threshold. Use for why-do-we-do-this and why-did-we-pick-that questions. Use how for runtime behavior.
unslop
Cut AI tells from any writing. Must always apply.
typescript-best-practices
TypeScript best practices. Use when reading or editing any .ts or .tsx file.
technical-writing
Layered technical-writing standard: Diátaxis structure, Google developer style sentences, STE instruction rules, Global English syntax. Use for `technical-writing` or when writing or reviewing docs, RFCs, readmes, PR descriptions, or commit messages.
teach
Explain a body of work plainly so a person actually understands it. Runs the `how` and `why` skills and weaves what they find into one clear explanation. Use for 'teach me this', 'help me really understand X', 'explain this change or subsystem to me'.
tdd
Use only when the user explicitly asks for TDD, a failing test, or a regression test, OR when the bug has an obvious cheap local test target. Skip when the test path is unclear, expensive, integration-heavy, or not requested.