Skill: question-resolver

What I do

I manage the process of finding answers to technical and domain-specific questions. I ensure that every question is classified, systematically researched using appropriate tools, and documented once resolved.

When to use me

• When faced with an unknown API, library, or codebase pattern
• To resolve ambiguity in user requests or requirements
• During research spikes to understand a new technology
• To track "known unknowns" that need resolution before proceeding

Core principles

1. Classify first — Is it answerable now (documentation), through research (spikes/data), or unanswerable (requires stakeholder input)?
2. Structured investigation — Use a methodical approach: hypothesise, search, verify.
3. Gather evidence — Rely on documentation, code, or experimental results rather than hearsay.
4. Document the "why" — Once resolved, record the answer and the evidence that supports it.

Patterns & examples

Question Log Template:

Escalation Triggers:

• Stuck: 30+ minutes without a clear path forward → Escalate or shift approach.
• Ambiguous: Requirement contradicts existing system behaviour → Escalate to stakeholder.
• Contradictory: Documentation differs from actual code behaviour → Trust code, but verify why.

Anti-patterns to avoid

• ❌ Rabbit holing — Spending hours researching a low-priority question.
• ❌ The "I think" trap — Accepting a plausible answer without actual verification.
• ❌ Ignoring "known unknowns" — Proceeding with a plan while key questions remain unanswered.

KB Reference

~/vaults/baphled/3. Resources/Knowledge Base/AI Development System/Skills/Thinking-Analysis/Question Resolver.md

Related skills

• critical-thinking — Validating the answers found
• assumption-tracker — Identifying the questions that need to be asked
• knowledge-base — Searching for existing answers
• epistemic-rigor — Distinguishing between theories and facts