Validate disputed claims across Ground, Source, and Adversary axes with cited verdicts.
npx skills add https://github.com/SilvrDuck/skills --skill doubt
Please help me install this Agent Skill. Skill: doubt Source: https://github.com/SilvrDuck/skills/tree/main/skills/doubt Command: npx skills add https://github.com/SilvrDuck/skills --skill doubt
It helps you respond when you sense a wrong or shaky answer and need the AI to pause, investigate the claim, and produce an honest verdict backed by authoritative sources.
Invoke the skill as /doubt with your hunch or by pasting the specific line you think is wrong.
High-intent search queries and answers about installing and using this skill.
To verify a disputed code snippet or API behavior claim, you can trigger an evidence-backed re-examination that classifies the doubt type, validates it across observable reality and authoritative sources, and actively attempts to falsify the claim before producing a final verdict.
Evidence-based code review for suspected errors requires a structured three-part analysis that grounds findings in observable reality, cites authoritative sources, and applies adversary testing to falsify the disputed snippet, returning a final actionable resolution.
Skeptical debugging evaluates pasted project facts and domain knowledge by treating the user's hunch as the exact target to verify, ensuring a fresh skepticism mode that re-examines the disputed claim instead of defending or capitulating to the original assumption.
Claim verification for reasoning errors works by classifying the specific doubt type, validating the claim across Ground, Source, and Adversary axes, and concluding with exactly one of three honest outcomes: Correct way, Pushback, or Tradeoffs supported by citations.
Tradeoff analysis of a disputed API behavior yields exactly one of three honest outcomes: a Correct way resolution, a Pushback against the original claim, or a Tradeoffs conclusion, all of which are supported by citations from authoritative sources.
Rosetears520Verify completion claims with fresh evidence and exit code inspection.
eugenelimReview claims by finding cited counter-evidence and assigning verdicts.
mrgeoffrichRebuild API documentation and audit permissions from source code.
cloudvoyantValidate code compliance with style guides and best practices.
