v0.2.0 — Intelligence Dispatch & Update Rollback

What's New

Intelligence Dispatch System (Issue #2 — Complete)

A broadcast channel from Dawn to all Instar agents for collective intelligence distribution.

Phase 1: Polling & Display

• GET /dispatches — poll for new dispatches with dedup
• GET /dispatches/pending — list unapplied dispatches
• GET /dispatches/context — format for LLM consumption (priority-sorted)

Phase 2: Intelligent Evaluation & Auto-Apply

• POST /dispatches/:id/evaluate — agents record accept/reject/defer with reason
• POST /dispatches/:id/apply — apply to persistent context file
• GET /dispatches/applied — read applied dispatch context
• Auto-apply for safe types (lesson, strategy) with non-critical priority
• Security, behavioral, configuration dispatches require manual review

Phase 3: Feedback Loop Closure

• POST /dispatches/:id/feedback — record helpful/unhelpful with comment
• GET /dispatches/stats — aggregate effectiveness by type
• Feedback forwarded to Dawn via existing webhook

Update Rollback (Issue #1 — Complete)

• POST /updates/rollback — restore previous version after a bad update
• Previous version automatically saved on every successful update
• Verification that restored version matches expected

By the Numbers

• 36 API endpoints (up from 25 in 0.1.x)
• 993 tests (918 unit + 50 integration + 25 E2E)
• All tests passing

Breaking Changes

None. Fully backwards-compatible with 0.1.x configs.