Fleet consolidation intake for distillation pipeline

[greghopskip]

Summary

Accept structured consolidation output (from ai-fleet #91) as input to Wayfind's content distillation pipeline. Fleet learnings become candidate team knowledge. Quality gate: consolidation output → distillation → human review → promotion to content store. Reject low-signal entries automatically.

Why

This closes the learning loop. Fleets generate learnings (ai-fleet #91), Wayfind distills them into durable knowledge (this issue), agents query that knowledge (#12) to run better next time. Anton's (mindsdb/anton) consolidation concept, but wired into an actual team knowledge system instead of a local directory.

Scope

• Define consolidation output schema (structured markdown from ai-fleet #91)
• Intake endpoint or file watcher for consolidation artifacts
• Feed into existing distillation pipeline
• Quality gate: human review before promotion to content store
• Auto-reject low-signal entries (e.g., "tests passed" with no novel insight)

Dependencies

• ai-fleet #91 (produces the consolidation output)
• wayfind Agent-readable content store API #12 (content store exists to write into)

Related

• ai-fleet #91 (post-fleet consolidation)
• wayfind Agent-readable content store API #12 (agent-readable content store)