Identify weaknesses and edge cases in designs through adversarial questioning.
npx skills add https://github.com/mcouthon/agents --skill critic-mcouthon
Please help me install this Agent Skill. Skill: critic Source: https://github.com/mcouthon/agents/tree/main/.github/skills/critic Command: npx skills add https://github.com/mcouthon/agents --skill critic-mcouthon
The Critic exists to challenge assumptions and identify edge cases, helping teams avoid flawed reasoning and weak designs by asking focused questions and probing until root causes are revealed.
Activate critic mode and begin probing assumptions, edge cases, and potential failure points.
High-intent search queries and answers about installing and using this skill.
Adversarial critique identifies edge cases and weaknesses in designs through focused questioning. It applies structured patterns like the 5 Whys to probe assumptions, evaluate failure modes, and reveal root causes during design reviews and strategy planning.
The best way to conduct a design review using adversarial thinking is to apply guided probing patterns that challenge assumptions. By systematically asking focused questions and exploring potential failure modes, teams can surface hidden risks and strengthen their overall design robustness.
Yes, you can use this approach to find root causes during requirements debates by applying the 5 Whys technique. It drills down through adversarial questioning to reveal underlying assumptions and root insights, ensuring robust reasoning in strategy planning and requirements analysis.
No, the critic approach does not provide direct solutions for risk analysis. It operates in a read-only mode that focuses entirely on asking probing questions and evaluating safety considerations to help teams uncover hidden risks and strengthen their own designs.
You should use structured probing patterns for quality assurance when you need to evaluate edge cases and failure modes in a design. Adversarial questioning helps teams avoid flawed reasoning by systematically challenging assumptions until root causes are fully revealed.
FrantssConduct adversarial reviews of project plans and design documents.
prathamchopra001Identify root causes of errors in Python and JavaScript code snippets and implement targeted corrections.
kagentiOrchestrate adversarial design reviews with simulated expert critiques and human feedback.
ulpi-ioAudit a repository and produce a GO, NO-GO, or GO-WITH-FIXES release verdict.
