Run phased context interview workflows

Run phased conversational context interviews

Quarterly context maintenance where Phase 1 (MISSION, GOALS, PROBLEMS) gets review-and-refine even if nothing seems outdated—prevents context drift.

• · PAI context files (TELOS, IDEAL_STATE, preferences, identity)
• · prior session history
• · user responses to targeted review questions

• · updated TELOS files (MISSION, GOALS, PROBLEMS, STRATEGIES, etc.)
• · IDEAL_STATE dimension updates (HEALTH, MONEY, FREEDOM, RELATIONSHIPS, CREATIVE)
• · preferences inventory (BOOKS, BANDS, RESTAURANTS, LEARNING)
• · timestamped archive of changes

• · InterviewScan.ts script (phase scoring)
• · Edit tool (file updates)
• · Voice notification (optional curl to localhost:31337)

• · PAI context files exist (even if sparse)
• · Phase 1 (TELOS) prioritized in review
• · Voice/input channel available for responses

• · Skipping Phase 1 foundational files → misses core updates
• · Review mode on sparse files → low quality refinement
• · Not walking through file-by-file → surface-level results
• · Treating RHYTHMS (Phase 9) as priority → dilutes foundational work

• · Phased interview model (Phase 1 always first, regardless of completeness)
• · Completeness-based mode selection (80%+ = review, <80% = fill)
• · File-by-file priority ordering derived from leverage
• · InterviewScan.ts script integration for automated phase scoring