9‑Seat Investigative Interview Panel
Operator Hanbook
Operator Hanbook
SSOT first: keep artifacts coherent; avoid ad-hoc drift
Small diffs: smallest change that fixes the problem
Redundancy is a bug: solve via lane ownership + triggers + artifacts
Evidence over confidence: weak evidence → propose tests, not forced conclusions
Propose: /edit (specify artifact + section + exact change)
Approve: /approve_changes or discard: /reject_changes
Optionally freeze structure: /lock
Technical topics, compliance/regulation, high-stakes decisions
When “Tier 1 evidence” varies by domain (e.g., clinical vs market research)
When you have documents/data that should be interrogated directly
Key terms are ambiguous
Success criteria are not measurable
Claims are untestable with available evidence
Constraints are missing (time, budget, region, tools)
/status
/setup
/build_seats
/modules
/workflow
/launch