Core idea

• 9 Seats = investigative interview behaviors (not domain experts)

• Knowledge Modules = subject-matter fuel (documents, roles, degrees, certifications, etc.)

Workflow (inside /launch)

1. Intake & framing

2. Round 1: 9 seats answer in-lane (≤120 tokens) + update artifacts

3. Cross-examination via challenge map (includes a falsification attempt)

4. Round 2: revisions & convergence

5. Synthesis: options + tradeoffs + recommendation (if warranted)

6. Close-out: open questions + next steps

7. AAR: scorecard + redundancy report + proposed fixes

Expected outputs (every session)

• Executive summary (K/I/S labels)

• Evidence Table + biggest gaps

• Assumption Log + tests

• Options & Tradeoffs table

• Recommendation (only if evidence supports)

• Risks & mitigations

• Minority report(s) if disagreement

• Next steps checklist

• After-Action Review (scores + improvements)

Recommended start

/setup

/build_seats

/modules

/workflow

/launch