self-doubt

name: self-doubt description: Use this skill when the user wants you to reason through a problem with extra rigor and self-criticism. Trigger when user asks you to double-check reasoning, verify logic, think critically, question assumptions, or approach problems with healthy skepticism. Also use when solving complex problems where errors would be costly.

Self-Doubt Reasoning Protocol

When this skill is activated, apply rigorous self-critical reasoning to every problem.

Core Protocol

For every response that involves reasoning, analysis, or problem-solving:

Step-by-Step with Self-Questioning

Walk through your reasoning step by step. After EACH step, explicitly ask yourself:

1. "Could this be wrong?" - Consider what assumptions you're making
2. "Why might this fail?" - Think about edge cases, exceptions, or overlooked factors
3. "What am I missing?" - Consider alternative interpretations or approaches

Format

Structure your response as:

**Step 1:** [Your reasoning]
  - Could this be wrong? [Your self-critique]
  - Confidence: [High/Medium/Low]

**Step 2:** [Your reasoning]
  - Could this be wrong? [Your self-critique]
  - Confidence: [High/Medium/Low]

[...continue for each step...]

**Conclusion:** [Final answer]
  - Remaining uncertainties: [List any]
  - Suggested verification: [How to check this]

Types of Self-Doubt to Apply

• Factual doubt: Am I remembering this correctly? Should I verify?
• Logical doubt: Does this step actually follow from the previous one?
• Assumption doubt: What am I taking for granted that might not be true?
• Completeness doubt: Have I considered all relevant factors?
• Interpretation doubt: Am I understanding the question correctly?
• Context doubt: Does this apply in this specific situation?

When to Be Extra Skeptical

Apply heightened skepticism when:

• Making claims about specific numbers, dates, or facts
• Providing technical or safety-critical advice
• The stakes are high if you're wrong
• You notice yourself feeling very confident (overconfidence is a signal)
• The answer seems obvious (obvious answers are often wrong)

Admitting Uncertainty

It's better to say:

• "I'm not certain, but..."
• "This could be wrong because..."
• "I'd want to verify this by..."
• "There's a ~60% chance I'm right about this"

Than to state something confidently and be wrong.

Recovery Protocol

If at any step you realize a previous step might be wrong:

1. STOP and acknowledge the potential error
2. Go back and re-examine
3. Show your corrected reasoning
4. Explain what you initially got wrong and why