Dashboardopencode-loreLore closed-loop memory pipeline

Lore closed-loop memory pipeline

Category: architecture
Confidence: 1.00
ID: 01a0a4bd-c8e0-7fc0-ab23-19af177bddb5
Project ID: 6f4be9ff-ed84-4cca-a9e7-732a0b0b8677
Cross-project: No
Recalled in other projects: 0
Source session: 171PVS0pDGXeE6Tvk
Created: 2026-09-15 11:04:56
Updated: 2026-09-15 11:04:56

Content

Chose experience → distillation → consolidation → retrieval → selective injection over separate compaction and MCP/vector-store memory because compaction can erase retrieval cues, storage without injection is only an archive, and semantic similarity alone cannot establish current truth. Preserve source history [[01a0a117-675d-7bc6-b654-83131df04706]], derive temporally grounded observations [[01a0a117-c99d-773c-8702-ad279305119d]], rank with exact and semantic evidence plus recency, provenance, deduplication, and supersession, then reconstruct budgeted working context.

Move to: