Stats
Actions
Tags
From cognition
Use when someone is evaluating a person, explaining behavior, or making a decision with expressed high confidence — auditing three layers: how information is judged (confirmation bias, anchoring, sunk cost), how behavior is explained (attribution errors), and how well someone knows what they know (overconfidence, Dunning-Kruger). Triggers on: "am I being biased", "why did they do that", "I'm sure about this", "편향 점검", "확증 편향", "왜 이렇게 생각하는 걸까", "내 판단이 맞나". Best for: high-stakes decisions, evaluating people, diagnosing why a conclusion feels certain. Not for: logical fallacy detection (use fallacy-detector), evidence calibration (use epistemic-reasoner).
How this skill is triggered — by the user, by Claude, or both
Slash command
/cognition:bias-auditorThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Use when:**
Use when:
Not for:
Step 1 — Gather the judgment. Get the specific decision, belief, or evaluation as concretely as possible.
Step 2 — Scan Layer 1: Judgment biases.
| Bias | Signal |
|---|---|
| Confirmation bias | Sought confirming info, avoided disconfirming |
| Availability heuristic | Recent/vivid events overweighted |
| Anchoring | First piece of info dominating subsequent estimates |
| Sunk cost | Past investment justifying continued action |
| Optimism/planning fallacy | Timelines or outcomes overestimated |
| Framing effect | Conclusion changes with presentation format |
Step 3 — Scan Layer 2: Attribution errors.
Step 4 — Scan Layer 3: Metacognitive accuracy.
Step 5 — Deliver the audit. Flag only what's actually present.
편향 감사 결과 / Bias Audit Results:
판단 편향 / Judgment Biases:
[Bias name]: [Specific manifestation in this case]
Impact: [What this likely distorts in the conclusion]
귀인 오류 / Attribution Errors:
[Error type]: [Where it appears]
Alternative explanation: [Situational factors the current explanation ignores]
메타인지 정확도 / Metacognitive Accuracy:
Confidence level: [calibrated / overconfident / underconfident]
[Specific evidence for this assessment]
핵심 편향 / Primary Bias:
[The single most consequential bias — the one most likely driving a wrong conclusion]
| Claude | You |
|---|---|
| Scans all three bias layers systematically | Describe the decision or judgment concretely |
| Names only biases that are actually present | Confirm whether the named pattern resonates |
| Connects each bias to a concrete implication | Decide what to check or do differently before acting |
| Identifies the single most consequential bias | Apply the corrective to the actual decision |
fallacy-detector — when the issue is argument structure, not cognitive biasepistemic-reasoner — when the issue is calibrating confidence to evidenceassumption-extractor — when the issue is hidden premises rather than cognitive distortionnpx claudepluginhub newkayak12/claude-skills --plugin cognitionDiagnoses which cognitive biases are actively distorting thinking in a specific situation. Activates on queries about bias, objectivity, and decision clarity.
Surfaces 3-4 cognitive biases in prior conversation reasoning or Libertee methods (Six Hats, Debate). Maps biases to session moments and poses one uncomfortable question challenging conclusions. Use after synthesis.
Detects and removes cognitive biases from reasoning using Julia Galef's Scout Mindset framework. Provides reversal tests, scope sensitivity checks, status quo bias tests, confidence interval audits, and full bias audits.