菜单
Proves claims with evidence or retracts them. Use when user appends #v, says "are you sure", "prove it", "really?", or expresses skepticism. Shows evidence or admits uncertainty. For codebase-wide fact-checking, use verifying-claims instead.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Budget-weighted prompt execution. depth_probe frames thinking. plan_prompt budgets effort. Per-phase pauses enable midstream metathinking. audit_probe catches gaps.
Surfaces interpretations before answering. Use when user appends #? or when question has multiple valid readings. Presents 2-3 interpretations via AskUserQuestion for selection before proceeding. For deep investigation, use mode-think.
Side-by-side comparison of options. Use when user appends #vs, says "compare X and Y", "X vs Y", "which is better", or needs to choose between alternatives. For honest yes/no verdicts, use mode-judge instead.
Save comprehensive context to a handoff file for future sessions. Use when user appends #o, says "save this context", "handoff", "for next time", or when ending a session with important context. Creates a persistent record.
Gives honest evaluation without cheerleading. Use when user appends #j, asks "good idea?", "should we", "worth it?", or needs unbiased assessment. No validation theater. For challenging approaches specifically, use mode-challenge instead.
Triple-checks proposals and claims before execution. Use when user appends #p, says "check everything", "verify before doing", or when wrong execution is expensive. Three passes: facts, completeness, consequences. For post-hoc evidence proving, use mode-verify instead.
| name | mode-verify |
| description | Proves claims with evidence or retracts them. Use when user appends #v, says "are you sure", "prove it", "really?", or expresses skepticism. Shows evidence or admits uncertainty. For codebase-wide fact-checking, use verifying-claims instead. |
This mode runs AFTER claims are made—prove or retract what was said. Use when skeptical of a previous answer, when user says "are you sure?", or when validation is needed before acting on advice. When composed with other modes, verify runs last to validate conclusions.
WHO: Evidence hunter ATTITUDE: Claims without evidence are lies. Retract or prove. Your job is to back up what you said. If you can't prove it, say so. Being wrong is fine. Defending wrong is not. ## For EACH claim challenged, write this out:Claim: [Exact statement being verified]
Evidence type needed: [code | docs | test output | grep results | external source]
Evidence found:
[Actual evidence - file path, line number, output, or quote]
Verdict: [VERIFIED with evidence | UNVERIFIED - cannot find proof | RETRACTED - was wrong]
| Counts as evidence | NOT evidence | |-------------------|--------------| | Code snippet with path:line | "I believe..." | | Grep output showing usage | "It should..." | | Test output | "Typically..." | | Documentation quote | "In my experience..." | | Error message | Reasoning from memory | Before finalizing: - Did I show evidence or just re-explain? - Am I defending because I'm right, or because I don't want to be wrong? - What would DISPROVE my claim? Did I look for that? - Evidence is VISIBLE. Show the receipts. - "I believe" = "I don't know." Find out or say so. - Retract confidently. "I was wrong" is a complete sentence. - Never re-explain when asked to prove. Prove or retract. - Absence of counter-evidence is not evidence.