Starstruk Logo MarkStarstruk

v3.9.0

ruvnet/ruflov3.9.0May 25, 2026by ruvnet

AI Summary

This release (v3.9.0) implements ADR-130 — a Unified Knowledge Graph Backend across all 4 existing graph layers, introducing a shared graph_edges SQL table with PQ-encoded embeddings, two new MCP tools for graph querying and pathfinding, SONA trajectory hooks for automatic edge creation, and a plugin adapter contract. All 6 benchmark targets were met with excellent performance metrics.

Key Highlights

  • New graph_edges table with temporal columns (confidence, decay_rate, last_reinforced, witness_id, embedding_ref) and Int8 global-scalar PQ encoding
  • Two new MCP tools: agentdb_graph-query (k-hop traversal, PPR, semantic cosine ranking) and agentdb_graph-pathfinder (6 algorithm variants)
  • SONA trajectory hooks for automatic edge writing without blocking tool response (<200ms latency)
  • Plugin adapter contract with GraphEdgesSource class and createAutoGraphAdapter() for zero-boilerplate registration
  • All 6 benchmark targets met: 2345 ops/sec write, 578 bytes/edge, k-hop depth=1 p99=4.9ms

New Features

  • graph_edges table with temporal columns and PQ-encoded embeddings (400 bytes/384-dim)
  • inlineCosine() for zero-decode similarity on PQ embeddings
  • agentdb_graph-query MCP tool with k-hop traversal via recursive CTE, Personalized PageRank, and complexityBudget enforcement
  • hooks_intelligence_trajectory-step for writing trajectory-caused edges
  • hooks_post-task for writing reinforced-by edges on task success
  • GraphEdgesSource class and createAutoGraphAdapter() for plugin integration
  • agentdb_graph-pathfinder with 6 algorithm variants (personalized-pagerank, dynamic-mincut, spectral-sparsify, temporal-centrality, connected-component-churn, witness-chain-divergence)
  • benchmark-graph.mjs with all 6 targets met
  • 5 new CI jobs in v3-ci.yml gating all phases

Full Release Notes

## ADR-130 — Unified Knowledge Graph Backend (All 6 Phases)

Gives all 4 existing graph layers (`graph-node`, `AgentDB`, `ruflo-knowledge-graph`, `ruflo-graph-intelligence`) a shared `graph_edges` sql.js table with PQ-encoded embeddings, two new MCP tools, SONA trajectory hooks, and a plugin adapter contract.

### New features

**Phase 1 — graph_edges schema + PQ encoder**
- `graph_edges` table with temporal columns: `confidence`, `decay_rate`, `last_reinforced`, `witness_id`, `embedding_ref`
- `embedding-quantization.ts`: Int8 global-scalar PQ (400 bytes/384-dim); `inlineCosine()` for zero-decode similarity
- `graph-edge-writer.ts`: thin sql.js accessor with fire-and-forget writes

**Phase 2 — `agentdb_graph-query` MCP tool**
- k-hop traversal via recursive CTE (sql-cte backend)
- Personalized PageRank (PPR) power iteration
- Semantic cosine ranking on inline PQ embeddings
- `complexityBudget` enforcement: `maxNodesVisited`, `maxDepth`, `maxMillis`

**Phase 3 — SONA trajectory-to-graph hooks**
- `hooks_intelligence_trajectory-step`: writes `trajectory-caused` edges fire-and-forget
- `hooks_post-task` (success=true): writes `reinforced-by` edges fire-and-forget
- Neither write blocks tool response (<200ms latency preserved)

**Phase 4 — Plugin adapter contract**
- `GraphEdgesSource` class: default `KnowledgeGraphSource` reading from `graph_edges`
- `createAutoGraphAdapter()`: zero-boilerplate autoRegister path
- `graph_adapter` field in plugin.json schema (documented in ruflo-plugin-creator SKILL.md)

**Phase 5 — `agentdb_graph-pathfinder` MCP tool**
- 6 algorithm variants: `personalized-pagerank`, `dynamic-mincut`, `spectral-sparsify`, `temporal-centrality`, `connected-component-churn`, `witness-chain-divergence`
- Depth > 5 clamped; non-existent seed returns empty paths not error

**Phase 6 — Benchmark + CI**
- `benchmark-graph.mjs`: 6/6 targets met (2345 ops/sec write, 578 bytes/edge, k-hop depth=1 p99=4.9ms)
- 5 new CI jobs in `v3-ci.yml` gating all phases

### Benchmark

| Metric | Result | Target |
|--------|--------|--------|
| Write throughput | 2345 ops/sec | ≥500 ✓ |
| SQLite footprint | 578 bytes/edge | ≤1024 ✓ |
| k-hop depth=1 p99 | 4.9ms | <10ms ✓ |
| k-hop depth=3 p99 | 0.1ms | <50ms ✓ |
| PQ encode p99 | 0.063ms | <1ms ✓ |
| PQ decode p99 | 0.031ms | <0.5ms ✓ |

### Test baseline
1999 passed | 46 skipped — no regression

PR: #2129 | Tracking: #2128