whetstone_DSL/sprint96_plan.md

Sprint 96 Plan: Knowledge Graph and Semantic Memory Fabric

Context

Long-lived programs benefit from structured memory linking pairs, failures, policies, and outcomes. Sprint 96 builds semantic memory fabric.

---

Goals

1. Build knowledge graph across transpilation artifacts and decisions
2. Enable semantic retrieval for planning and review workflows
3. Improve pattern reuse and repeated issue prevention
4. Preserve traceability from strategy to outcome

---

Steps

Step 1189: Transpilation knowledge graph schema (12 tests)

Step 1190: Artifact-to-graph ingestion pipeline (10 tests)

Step 1191: Semantic retrieval API for decisions/failures/policies (10 tests)

Step 1192: Pattern recurrence detector from graph signals (10 tests)

Step 1193: Strategy recommendation hooks from graph context (10 tests)

Step 1194: Graph consistency and retention policy integration (8 tests)

Step 1195: whetstone_query_transpile_graph MCP tool (8 tests)

Step 1196: whetstone_get_related_migration_patterns MCP tool (8 tests)

Step 1197: Semantic memory report artifact (8 tests)

Step 1198: Sprint 96 integration summary + regression (8 tests)

---

Memory Rule

• Graph-derived recommendations must cite source artifacts and confidence.