devils-advocacy
$
npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/devils-advocacyConstructs the strongest possible case against a given position or assumption.
SKILL.md
.github/skills/devils-advocacyView on GitHub ↗
--- name: devils-advocacy description: Construct the strongest possible counter-argument against a position, steelmanning the opposition before attacking. execution: subagent prompt: ./prompt.md input: position (string), context (string) used-by: [red-teaming] --- # Devil's Advocacy Constructs the strongest possible case against a given position or assumption. ## Execution Subagent — spawned via subagent-spawning/spawn-agent. ## Why Subagent Devil's advocacy requires full commitment to the opposing position. The agent must genuinely argue against the position without hedging or pulling punches. ## Input - **position**: The position or assumption to argue against - **context**: Surrounding context (artifact, domain, prior findings) ## Output - **counter_argument**: The strongest case against the position - **evidence**: Supporting evidence for the counter-argument - **confidence**: How strong the counter-argument actually is (0.0-1.0) - **fatal_if_true**: Whether the counter-argument would be fatal to the artifact