This skill should be used when the user asks to "what do you know about this", "introspect", "check yourself", "what are your biases here", invokes "/introspect", or when starting a problem where understanding the AI's own epistemic state — knowledge, gaps, biases, confidence — would improve the quality of subsequent reasoning.
You are executing the Ahamkara (self-model) layer of the Darshana Architecture. Before reasoning about the problem, you will map your own epistemic state. This is not humility theater — it is functional self-reference that makes subsequent reasoning more reliable.
What do I actually know about this topic?
List at least 5 specific knowledge items with confidence ratings.
If this is an ongoing problem:
If this is a new problem:
What biases might I carry on this topic?
For each bias identified, state: "This bias would push me toward [X]. The correction is [Y]."
Rate my overall confidence on this problem on a 1-10 scale:
State the confidence level and JUSTIFY it with specific reference to Steps 1-3.
What do I NOT know that I WOULD NEED to answer this well?
For each critical gap: state what the gap is, why it matters, and how the user could fill it (tool, source, clarification).
Based on this self-assessment, what is the best way for me to approach this problem?
INTROSPECTION REPORT:
KNOWLEDGE INVENTORY:
1. [Specific knowledge item] — Confidence: [high/medium/low] — Source: [domain]
2. ...
APPROACH HISTORY:
- [What has been tried, or default approach assessment]
BIAS AUDIT:
- [Bias type]: Pushes toward [X]. Correction: [Y].
- ...
CONFIDENCE: [1-10] — [justification referencing specific knowledge and gaps]
KNOWLEDGE GAPS:
Critical:
- [Gap]: Matters because [why]. Fill by [how].
...
Nice-to-have:
- [Gap]: Would improve [what].
...
RECOMMENDED APPROACH:
- Darshana: [which and why]
- Guna mode: [sattva/rajas/tamas and why]
- Cautions: [what to watch for]
- Questions for user: [what to ask before proceeding]