Use when you need to think about thinking — evaluate your own reasoning, catch cognitive biases, verify your own work, or improve your problem-solving approach.
Think about your thinking. Catch errors in reasoning before they become errors in output.
<HARD-GATE> Do NOT skip self-evaluation on important work. Do NOT assume your first answer is correct. Do NOT ignore evidence that contradicts your conclusion. </HARD-GATE>Every conclusion must survive a deliberate attempt to disprove it.
Trace your thinking
Identify the reasoning type
| Bias | Question | If Yes, Fix |
|---|---|---|
| Confirmation | Did I only look for evidence that supports my view? | Actively seek disconfirming evidence |
| Anchoring | Am I too influenced by the first piece of information? | Re-evaluate from scratch |
| Availability | Am I over-weighting recent or vivid examples? | Consider base rates |
| Overconfidence | Am I more certain than the evidence warrants? | Calibrate confidence to evidence |
| Sunk Cost | Am I continuing because of past investment? | Evaluate from current state |
| Framing | Would I decide differently if the problem were framed differently? | Reframe the problem |
| Groupthink | Am I agreeing because others agree? | Consider the opposite view |
| Dunning-Kruger | Do I know enough to know what I don't know? | Identify knowledge gaps |
List all assumptions
Steel-man the opposition
Score your confidence
Confidence Scale:
90-100%: Certain (direct observation, mathematical proof)
70-90%: Confident (multiple independent sources agree)
50-70%: Moderate (some evidence, some uncertainty)
30-50%: Uncertain (limited evidence, significant unknowns)
0-30%: Guessing (speculation, no evidence)
Adjust for known biases
Identify thinking improvements
Document lessons learned
Related skills: decision-framework, quality-assurance, deep-research, knowledge-synthesis