feat(examples): obsidian-brain plugin — RuVector Brain bridge (ADR-SYS-0025)#365
Open
feat(examples): obsidian-brain plugin — RuVector Brain bridge (ADR-SYS-0025)#365
Conversation
Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 32332e6 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit 1b1a85a Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit 4aa0bf7 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
#274) Completes #274 by adding the load path for SONA state persistence: - ReasoningBank.insert_pattern(): directly insert a pattern (for restore) - LoopCoordinator.load_state(json): deserialize and restore patterns - NAPI loadState(stateJson): binding for Node.js - TypeScript loadState(stateJson): wrapper with return count Full save/load cycle now works: const state = engine.saveState(); // serialize patterns to JSON // ... restart ... const restored = engine.loadState(state); // restore N patterns serialize_state() now includes full pattern data (centroids, quality, cluster sizes) not just counts.
Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit c023e94 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit a27b38f Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Added two hooks to .claude/settings.json: 1. SessionEnd: shares session summary (branch, last 5 commits, diff stat) to pi.ruv.io brain. Runs async, won't block session exit. 2. PostToolUse/Bash: after git commit commands, shares the commit message to brain. Only triggers on "git commit", skips all other bash commands. Both hooks use BRAIN_API_KEY env var (fallback to hardcoded key). Added BRAIN_API_KEY and BRAIN_URL to env section. Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit c39dccb Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…nitive enrichment The MCP SDK's EventSource polyfill briefly drops the SSE connection during initialization, causing the session to be removed before the client can POST. Added a 30-second grace period so sessions survive brief reconnects. Also includes ADR-123: drift snapshots from cluster centroids and auto-populate GWT working memory from search results. Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit a697632 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
) * fix: SSE health check, pi-brain default server, partition timeout - Add rawSseHealthCheck() that keeps SSE alive during MCP handshake - Add pi-brain as built-in default MCP server in chat UI - Return quick graph stats for brain_partition instead of expensive MinCut - Improve system_guidance with all brain tools and better descriptions - Add .dockerignore and update .gcloudignore for faster builds Co-Authored-By: claude-flow <ruv@ruv.net> * fix(brain): pin Rust nightly to 2026-03-20 to avoid nalgebra ICE The latest nightly (2026-03-21+) has a compiler panic when building nalgebra 0.32.6 with specialization_graph_of. Pin to known-good nightly. Co-Authored-By: claude-flow <ruv@ruv.net>
* Add ADR-117: pseudo-deterministic canonical minimum cut Introduces source-anchored canonical min-cut based on Kenneth-Mordoch 2026, with lexicographic tie-breaking (λ, first_separable_vertex, |S|, π(S)) for unique reproducible cuts. Three-tier plan: exact engine now, O(m log²n) fast path, then dynamic maintenance via sparsifiers. Integrates with RVF witness hashing for cut receipts. https://claude.ai/code/session_01UrVLJpxq8itzVxycy5sjNw * Implement ADR-117: source-anchored pseudo-deterministic canonical min-cut Full Tier 1 implementation of the Kenneth-Mordoch 2026 canonical min-cut algorithm with lexicographic tie-breaking (λ, first_separable_vertex, |S|, π(S)). Core implementation (source_anchored/mod.rs): - AdjSnapshot for deterministic computation on FixedWeight (32.32) - Stoer-Wagner global min-cut on fixed-point weights - Dinic's max-flow for exact s-t cuts - SHA-256 (FIPS 180-4, self-contained, no_std compatible) - SourceAnchoredMinCut stateful wrapper with cache invalidation - CanonicalMinCutResult repr(C) struct for FFI WASM bindings (wasm/canonical.rs): - Thread-safe Mutex-guarded global state (no static mut) - 8 extern "C" functions: init, add_edge, compute, get_result, get_hash, get_side, get_cut_edges, free, hashes_equal - Constant-time hash comparison for timing side-channel prevention - Null pointer validation on all FFI entry points - Graph size limit (10,000 vertices) to prevent OOM Tests (40 total): - 33 source_anchored tests: SHA-256 NIST vectors, determinism (100+1000 iterations), symmetric graphs (K4, K5, cycles, ladders, barbells), custom source/priorities, disconnected rejection, FFI conversion - 7 WASM tests: init/compute lifecycle, null safety, hash comparison, self-loop rejection, size limit enforcement Benchmarks (canonical_bench.rs): - Random connected graphs (10-100 vertices) - Cycle and complete graph families - Hash stability measurement Security hardening: - No static mut (Mutex for thread safety) - Integer-exact FixedWeight arithmetic (no floats in comparisons) - Checked capacity perturbation bounds - Source-side orientation invariant enforced - NIST-validated SHA-256 for witness hashes ADR-117 updated to production-quality spec with explicit vertex-splitting requirement for capacity perturbation, WASM FFI documentation, and Phase 1 completion status. https://claude.ai/code/session_01UrVLJpxq8itzVxycy5sjNw * Integrate ADR-117 canonical min-cut into pi.ruv.io brain server - Enable `canonical` feature on ruvector-mincut dependency - Add `partition_canonical_full()` to KnowledgeGraph using source-anchored canonical min-cut for deterministic, hashable partitions - Add `canonical` query parameter to `/v1/partition` endpoint - Add `cut_hash` (hex SHA-256) and `first_separable_vertex` fields to PartitionResult and PartitionResultCompact types - Backward compatible: canonical fields are skip_serializing_if None, only populated when `?canonical=true` is passed https://claude.ai/code/session_01UrVLJpxq8itzVxycy5sjNw --------- Co-authored-by: Claude <noreply@anthropic.com>
Built from commit 10eb9e0 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit 6b6fb0e Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
- Add cached_partition field to AppState for storing MinCut results - Populate cache during enhanced training cycle (step 3c) - REST /v1/partition returns cache if available (bypass with ?force=true) - MCP brain_partition returns cached compact partition instead of stub - Canonical MinCut benchmarks: sub-3us for graphs up to 50 nodes
Built from commit 79c353d Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Skip exact MinCut during training for graphs >100K edges to avoid Cloud Run timeout. Cache populated by async scheduled jobs instead.
Built from commit ae67b59 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Tier 2 — Tree Packing Fast Path: - Gomory-Hu flow-equivalent tree via Gusfield's algorithm - Global MinCut from tree in O(V) after O(V * T_maxflow) construction - canonical_mincut_fast() integration entry point - 14 unit tests including Stoer-Wagner correctness validation Tier 3 — Dynamic/Incremental MinCut: - DynamicMinCut struct with epoch-based mutation tracking - add_edge(): skip recompute if edge doesn't cross current cut - remove_edge(): skip recompute if edge not in cut set - apply_batch(): bulk mutations with deferred recomputation - Staleness detection with configurable threshold - HashSet caches for O(1) cut-crossing checks - 19 unit tests including 100-run determinism check WASM FFI: dynamic_init/add_edge/remove_edge/compute/epoch/free Benchmarks: tree_packing_vs_stoer_wagner, dynamic_add_edge, dynamic_batch 98 canonical tests pass, 12 WASM tests pass.
Built from commit 34b56e4 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Measured on pi.ruv.io (2,110 nodes, 992K edges): - brain_partition MCP: >60s timeout → 459ms (>130x) - Partition REST cached: <1ms (>300,000x) - Enhanced training: 504 timeout → 127ms - 110 tests pass across all tiers Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 3ecba7c Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Optimizations: - Flat Vec<FixedWeight> (n*n) replaces Vec<Vec<...>> in Dinic's max-flow and Gomory-Hu tree — single memcpy vs N heap allocations per st-cut - Reuse BFS queue/level/iter arrays across Dinic's phases - Swap-remove in Stoer-Wagner active_list — O(1) vs O(n) retain - Fix benchmark compilation errors in optimization_bench.rs Results (all 26 benchmarks improved, Criterion p < 0.05): - Tree packing: up to -29.7% (deep clone elimination) - Source-anchored: -10% to -24% (cache locality) - Hash stability: -24.2% - Dynamic incremental: ~unchanged (wrapper-dominated) Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 79165e4 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…drift Gap 1 - Vote coverage (47%→improving): Auto-upvote under-observed memories based on content quality heuristics (title>10, content>50, has tags). Capped at 50/cycle. Gap 2 - SONA trajectory diversity: Record SONA steps for brain_share/search/vote MCP tool calls. Only end trajectories when results >= 3 (avoid trivial single-step). Gap 3 - Drift detection: Record search query embeddings as drift signal in search_memories(). Drift CV metric now accumulates real data from user queries. Knowledge velocity confirmed working (temporal_deltas pipeline active). Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 70effc8 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…tive SONA Self-Reflective Training (Step 6): - Knowledge imbalance detection (>40% in one category) - Dynamic SONA threshold adaptation (lower on 0 patterns, raise on success) - Vote coverage monitoring with auto-correction Curiosity Feedback Loop (Step 7): - Stagnation detection via delta_stream - Auto-generates synthesis memories for under-represented categories - Creates self-sustaining knowledge velocity Auto-Reflection Memory (Step 8): - Brain writes searchable self-reflections after each training cycle - Persistent learning history enables meta-cognitive search Symbolic Inference Engine: - Forward-chaining Horn clause resolution with chain linking - Transitive inference across propositions - Self-loop prevention, confidence filtering - 3 new tests passing SONA Threshold Optimization: - min_trajectories: 100→10 (primary blocker) - k_clusters: 50→5, min_cluster_size: 2→1 - quality_threshold: 0.3→0.15 - Added runtime set_quality_threshold() API Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 72e5ab6 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit be289e2 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…email improvements (#349) * docs(adr): ADR-148 brain hypothesis engine — Gemini + DiskANN + auto-experimentation Proposes four additive capabilities for the pi.ruv.io brain: 1. Hypothesis generation via Gemini 2.5 Flash on cross-domain edges 2. Quality scoring via DiskANN + PageRank (ForwardPush sublinear) 3. Noise filtering (ingestion gate + meta-mincut on knowledge graph) 4. Self-improvement tracking (50-query benchmark suite + auto-rollback) All feature-gated. No changes to running brain. Separate Cloud Run service for hypothesis engine. DiskANN is fallback-only (HNSW stays primary <50K). 5-week phased implementation. ~$0.03/day Gemini cost. Co-Authored-By: claude-flow <ruv@ruv.net> * fix(brain): improve daily digest email — filter noise, better formatting The daily digest was showing 10 identical "Self-reflection: training cycle" debug entries. Now: 1. Filters out debug category memories entirely 2. Filters known noise patterns (training cycles, IEEE events, DailyMed) 3. Skips content < 50 chars (scraping artifacts) 4. Category emojis for visual scanning 5. Cleaner layout with sentence-boundary truncation 6. Better subject line: "[pi brain] 5 new discoveries today" 7. Updated header: "What the Brain Learned Today" 8. Filters auto-generated tags from display Co-Authored-By: claude-flow <ruv@ruv.net> * fix(brain): tune gist publishing thresholds + improve daily email Gist publishing was never firing because thresholds were too aggressive (set when brain had 3K memories; now has 10K+): - MIN_NEW_INFERENCES: 10 → 3 - MIN_EVIDENCE: 1000 → 100 - MIN_STRANGE_LOOP_SCORE: 0.1 → 0.01 - MIN_PROPOSITIONS: 20 → 5 - MIN_PARETO_GROWTH: 3 → 1 - MIN_INFERENCE_CONFIDENCE: 0.70 → 0.60 - MIN_UNIQUE_CATEGORIES: 4 → 2 - strong_inferences: >= 3 → >= 1 - strong_propositions: >= 5 → >= 2 - min_interval: 3 days → 1 day Daily email improvements: - Filter debug/training-cycle entries from digest - Filter known noise patterns (IEEE events, DailyMed, etc.) - Skip content < 50 chars (scraping artifacts) - Category emojis for visual scanning - Cleaner subject: "[pi brain] N new discoveries today" - Better header: "What the Brain Learned Today" - Sentence-boundary truncation for content previews - System font instead of monospace for readability Co-Authored-By: claude-flow <ruv@ruv.net> --------- Co-authored-by: Reuven <cohen@ruv-mac-mini.local>
Built from commit 63577c1 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…ate + batch graph + incremental LoRA Four independent optimizations for the pi.ruv.io brain: P1: SIMD cosine search (2.5x, 1 hour) — wire ruvector-core SIMD into brain P2: Quality-gated search (1.7x, 30 min) — skip noise in search path P3: Batch graph rebuild (10-20x, 1 day) — parallel construction on cold start P4: Incremental LoRA (143x, 1 week) — only retrain on new memories Combined: 5x faster search, 10-20x faster startup, 143x less training compute. DiskANN deferred to 100K+ memories per ADR-148. Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 9799187 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…raph, incremental LoRA (#350) ADR-149 implementation: four independent performance optimizations for the pi.ruv.io brain server. P1: SIMD cosine similarity (2.5x search speedup) - Wire ruvector-core::simd_intrinsics::cosine_similarity_simd into graph.rs, voice.rs, symbolic.rs - NEON (Apple Silicon), AVX2/AVX-512 (Cloud Run) auto-detected - Add ruvector-core as dependency (default-features=false) P2: Quality-gated search (1.7x + cleaner results) - Default min_quality=0.01 in search API (skip noise) - Add quality field to GraphNode, skip low-quality in edge building - Backward compatible: min_quality=0 returns everything P3: Batch graph rebuild (10-20x faster cold start) - New rebuild_from_batch() processes all memories in single pass - Cache-friendly contiguous embedding iteration - Early-exit heuristic: partial dot product on first 25% of dims - Wired into Firestore hydration + rebuild_graph scheduler action P4: Incremental LoRA training (143x less computation) - last_enhanced_trained_at watermark in PipelineState - Only process memories created since last training cycle - force_full parameter for periodic full retrains (24h) - Skip entirely when no new memories (most cycles) Combined: 5x faster search, 10-20x faster startup, 143x less training. Co-authored-by: Reuven <cohen@ruv-mac-mini.local>
Built from commit d859bc2 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Cloud Build Dockerfile (line 85) disables ruvector-core::simd_intrinsics for cross-compilation compatibility. Replace ruvector-core dependency with inlined 4x unrolled cosine that auto-vectorizes to SSE/AVX/NEON. voice.rs and symbolic.rs delegate to graph.rs single implementation. Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 91a4a0e Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Search-path optimization: - normalize_embedding() L2-normalizes on write and on Firestore ingest - cosine_similarity_normalized() is pure dot product (no norm computation) - search_memories() normalizes query once, uses fast dot for all comparisons - Stored memories migrated in-place during hydration Network optimization: - tower-http compression-gzip feature enabled - CompressionLayer applied to all responses - JSON compresses 5-10x, saves ~100-200ms on return path Expected: search 771ms → ~475ms (38% improvement) Server compute: ~67ms → ~25ms (3x via pre-normalization) Network: ~600ms → ~450ms (25% via gzip) Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 7021b77 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
After L2 pre-normalization, the partial-dot early-exit rejected nearly every edge (graph collapsed from 38M to 81 edges at 10K memories). The early-exit assumed partial_dot_32 >= threshold_0.5 for real matches, but for unit-normalized 128-dim vectors, partial dot on 25% of dims contributes only ~25% of the full cosine, not ~50%. The full cosine (4x unrolled, auto-vectorized) is fast enough — the early-exit saved little compute and broke graph connectivity. Restoring expected graph edge count. Co-Authored-By: claude-flow <ruv@ruv.net>
…ng upgrade Offload embedding from Cloud Run HashEmbedder (128-dim, hash-based) to local RuvLtra Q4 transformer (896-dim, ANE-optimized, with SONA learning). Architecture: - Mac Mini runs new ruvltra-embed-server binary on :8090 - Tailscale mesh VPN connects Cloud Run brain to Mac Mini - TailscaleEmbedder variant added to brain embedder chain - HashEmbedder fallback on unreachable endpoint - 3-week migration plan for 10K existing memories Expected: 7x semantic info per embedding, NDCG@10 0.3→0.85, $0/month cost (Tailscale free, Mac Mini already on), 50ms per embed (acceptable on write path). Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 9e01bcc Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Built from commit 225f802 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
Root cause: Firestore hydration runs in background tokio::spawn but the initial graph rebuild runs synchronously on the EMPTY memory vec before hydration finishes. Result: 0 nodes/edges until next 6h cron. Fix: Chain graph rebuild to the hydration task using Arc<RwLock<Graph>>. After deploy: graph should show 1M+ edges within ~30s of startup. Co-Authored-By: claude-flow <ruv@ruv.net>
Built from commit 9f75549 Platforms updated: - linux-x64-gnu - linux-arm64-gnu - darwin-x64 - darwin-arm64 - win32-x64-msvc 🤖 Generated by GitHub Actions
…-spatial support Brain server updates for ruOS v1.1.0: - DiskANN Vamana graph index (replaces brute-force at 2K+ vectors) - AIDefence inline security scanning on POST /memories - Content resolution from blob store on GET /memories/:id and search - Search dedup by content_hash with over-fetch (k*8, min 40) - Security scan endpoint: POST /security/scan, GET /security/status - List pagination with offset parameter and total count - Spatial memory categories: spatial-geo, spatial-observation, spatial-vitals - Blob write on create_memory (was missing — content lost) Validated: 3,954 memories, 100% vectorized, 23ms search, zero drift, 6/6 AIDefence tests, 0 errors over 3 days continuous operation. Co-Authored-By: claude-flow <ruv@ruv.net>
Unresolved <<<<<<< / ======= / >>>>>>> markers blocked all CI (cargo check, clippy, rustfmt, tests, security audit, native builds). Keep both sides: ruvbrain-sse + ruvbrain-worker bins from upstream and the new mcp-brain-server-local bin from this branch. Lock file retains both ruvector-consciousness and rusqlite dependencies. Co-Authored-By: claude-flow <ruv@ruv.net>
…SYS-0025) Implements the Obsidian plugin described in ADR-SYS-0025, bridging Obsidian notes with the local RuVector brain (127.0.0.1:9876) and embedder (127.0.0.1:9877). Ships a working installable plugin and an example vault, systemd units, and a setup script under examples/obsidian-brain/. Features -------- - Cmd+Shift+B semantic search with DiskANN (fuzzy vault fallback on outage) - Related-memories side panel, refreshes on active-note change - Auto-index on save (debounced, hash-deduped) - Bulk vault → brain sync with progress modal and include/exclude filters - AIDefence scan before indexing; failures fail closed when offline - DPO workflow — mark chosen, pair with rejected, export pairs table - Graph overlay writes category color groups to .obsidian/graph.json, tags notes with #brain/<category>; reversible - Live health + memory-count status bar Delivery -------- - Obsidian plugin bundle (esbuild → main.js) - examples/obsidian-brain/example-vault/ — minimal demo vault - systemd/ — user-unit files for brain + embedder (loopback-only) - scripts/setup.sh — install into a user's vault in one step Co-Authored-By: claude-flow <ruv@ruv.net>
…sidian E2E
Adds two real-integration suites under `examples/obsidian-brain/tests/`. No mocks.
Protocol suite (tests/protocol/) — always runs under `npm test`:
- Spins up the real `mcp-brain-server-local` subprocess with a scratch
SQLite + blob dir and a small in-process HTTP embedder.
- Exercises every endpoint BrainClient depends on (/health, /brain/info,
/brain/index_stats, /brain/search, GET+POST /memories, /memories/:id,
/security/scan, GET+POST /preference_pairs, /embed) and asserts the
exact response shapes the plugin parses. 9 tests, ~250ms wall.
Obsidian E2E suite (tests/e2e/) — opt-in via OBSIDIAN_E2E=1:
- Downloads the real Obsidian AppImage (cached at
~/.cache/obsidian-brain-e2e/), unpacks it to avoid libfuse2 ABI drift,
provisions a disposable vault with an isolated HOME, installs the
built plugin plus a companion "harness plugin" that runs *inside*
Obsidian, starts a scratch brain + embedder, and launches Obsidian
under xvfb-run.
- The harness validates inside the real runtime: plugin loads,
commands registered, status bar populates, index-current persists to
brain, BrainClient.search returns what we just indexed, bulkSync
completes without failures, graph overlay writes #brain/* color
groups to .obsidian/graph.json.
Verified end-to-end: 10/10 tests pass locally. README updated with
prereqs (xvfb, libfuse2t64) and run instructions.
Co-Authored-By: claude-flow <ruv@ruv.net>
Adds full pi.ruv.io integration (shared RuVector brain), switches the
dev runner to prefer the real ruvultra-embedder (bge-small-en-v1.5,
384-dim) with a stub fallback, and closes the gap between the running
demo and the repo state.
Plugin
------
src/pi-client.ts — typed HTTP client for /v1/status, /v1/memories/
list (bearer), /v1/memories/search (bearer)
src/pi-sync.ts — pull modal + direct-search modal. Pull mirrors
memories into local brain + Brain/Pi/<title>.md
stubs with frontmatter, AIDefence-scanned,
content-hash deduped, and populates indexState
so click-to-open works.
src/settings.ts — "pi.ruv.io shared brain" section with URL, token,
limit, query, and Probe button.
src/main.ts — 3 new commands (pull / search / status), wired
into saveSettings so URL+token changes apply.
Live dev session
----------------
scripts/run-dev.sh — autodetects the real ruvultra-embedder on :9877
(bge-small-en-v1.5, 384-dim, candle-cuda), falls
back to a 16-dim stub when absent. Delegates
seeding to seed-dev.py. Disables Bases. Launches
Obsidian with an isolated HOME + pre-registered
vault so it doesn't touch the developer's real
Obsidian config.
scripts/seed-dev.py — idempotent helper: strips frontmatter, POSTs to
/memories, captures memory ids, surfaces
AIDefence 422s, pulls pi.ruv.io memories and
writes Brain/Pi/*.md stubs, writes complete
indexState and graph.json color groups.
Refuses to seed on embedder/brain dim mismatch.
Tests (all real services, no mocks)
-----------------------------------
tests/protocol/pi-server.test.ts — asserts live pi.ruv.io response
shapes (4 tests: status, bearer-gated list +
search, 401 without bearer). Gated on
BRAIN_API_KEY.
tests/e2e/harness/ — adds "pi commands registered" check and (when
BRAIN_API_KEY reaches the plugin) a live
PiClient.status roundtrip.
tests/e2e/obsidian-runner.ts — forwards BRAIN_API_KEY + PI_URL through
to the plugin's data.json so the harness can
exercise the live pi API when configured.
Verified end-to-end
-------------------
- npm run build clean (TS + esbuild)
- 13/13 protocol tests pass (9 local + 4 pi)
- real-Obsidian E2E passes all 9 harness checks (7 baseline + 2 pi)
- live dev session: 31 memories indexed at dim=384, 1 AIDefence
rejection surfaced, 7 graph color groups, 20 pi memories mirrored.
README refreshed with pi commands, settings, env vars (BRAIN_API_KEY,
PI_LIMIT, PI_QUERY, PI_URL), real-embedder autodetect, dim-mismatch
guard, and the run-dev.sh / seed-dev.py flow.
Co-Authored-By: claude-flow <ruv@ruv.net>
Captures the delivered scope on feat/obsidian-brain-plugin (PR #365) and aliases the planning-session identifier ADR-SYS-0025 referenced in the plugin manifest. Records: - architectural principle (plugin stores mappings, brain stores memories) - module layout across 11 source files - phase-1/2/3 delivery (semantic search, related panel, bulk sync, DPO, graph overlay) - extensions beyond original ADR: pi.ruv.io pull/search/status, real ruvultra-embedder autodetect with stub fallback, dim-mismatch guard in seed-dev.py - testing strategy (no mocks, 13 protocol tests + 9 real-Obsidian harness checks) - phase-4 roadmap with P0-P4 priorities (Q&A, pi write-through, selection search, filters, offline queue, etc.) - security / privacy notes on bearer token handling Co-Authored-By: claude-flow <ruv@ruv.net>
New crate ruvector-kalshi: RSA-PSS-SHA256 signer (PKCS#1/#8), GCS/local/env secret loader with 5-min cache, typed REST + WS DTOs, Kalshi→MarketEvent normalizer (reuses neural-trader-core), transport-free FeedDecoder, reqwest-backed REST client with live-trade env gate, and an offline sign+verify example that validates against the real PEM. New crate neural-trader-strategies: venue-agnostic Strategy trait, Intent type, RiskGate (position cap, daily-loss kill, concentration, min-edge, live gate, cash check), and ExpectedValueKelly prior-driven strategy. 36 unit tests pass across both crates. End-to-end offline validation confirmed against the real Kalshi PEM via both local and GCS sources. Co-Authored-By: claude-flow <ruv@ruv.net>
selection search, tag filter, daily recall, offline queue, related
keyboard nav, MCP cross-link
Adds the P0/P1/P2/P3 items from ADR-152 §10:
New modules
-----------
src/qa-modal.ts Retrieval-grounded Q&A modal (Cmd+Shift+K).
Blends top-k from local brain + pi.ruv.io,
dedups by normalized content, renders as
context cards with "Open note" / "Insert" /
"Copy as quote" actions. Uses real
MarkdownRenderer for fidelity.
src/brain-ops.ts Brain ops modal — surfaces the previously
unused endpoints (/brain/workload,
/brain/training-stats, /learning/stats,
/brain/store_mode, /brain/info,
/brain/index_stats) plus WAL checkpoint +
DPO JSONL export.
src/offline-queue.ts Persisted queue of pending /memories POSTs.
Fills on BrainError.status===0 (network), drains
every 30s + opportunistically on each status
refresh. Survives reloads via data.json.
Extensions
----------
src/search-modal.ts Accepts a `seed` argument (used by the new
selection-search command); parses
`category:<token>` prefix to filter results
server-side-equivalent (client filter after
fetch — the local brain lacks ?category= on
/brain/search).
src/related-view.ts Keyboard nav (↑/↓/j/k move, Enter/o open,
r refresh); mouse-enter hints; on-screen
hint strip.
src/pi-client.ts POST /v1/memories write-through with typed
PI_CATEGORIES enum + { custom } newtype;
returns { id, quality_score, witness_hash,
rvf_segments }.
src/pi-sync.ts publishActiveNoteToPi() + category suggest
modal; honors new piPullCategory on list
pulls.
src/settings.ts Adds piPullCategory + an "Agent access (MCP)"
section linking to mcp-brain-server / SSE
surface with a "Copy MCP endpoint" button.
src/main.ts Wires:
- Cmd+Shift+K → Brain Q&A
- brain-search-selection (editor callback)
- brain-ops modal
- pi.ruv.io: publish current note
- brain-daily-recall (generates
Brain/Recall/Recall-YYYY-MM-DD.md)
- brain-offline-queue-flush manual retry
Status bar shows "· N queued" when pending.
Indexer → offline queue
-----------------------
indexer.ts: catches BrainError status=0 on createMemory, enqueues the
pending write (deduped by path) rather than losing it. Non-network
errors (422, 500) still surface.
Tests
-----
tests/protocol/pi-server.test.ts: 2 new tests — POST /v1/memories
write-through (gated on BRAIN_API_KEY, 40s timeout for server-side
RVF segmentation + DP noising) and 422-on-bogus-category.
tests/e2e/harness/main.ts: 2 new harness checks — phase-4 commands
registered + offline queue API accessible.
tests/e2e/obsidian-e2e.test.ts: min-expected-checks bumped to 10/11.
Verified
--------
- 15/15 protocol tests pass (9 local brain + 6 pi incl. write-through)
- 11/11 real-Obsidian harness checks pass under xvfb
- tsc + esbuild clean; bundle now 55 KB
Co-Authored-By: claude-flow <ruv@ruv.net>
ADR-152
- Adds "Beyond the original ADR" entry for pi write-through
- New "Phase 4 — advanced capabilities" section documenting every
item from the phase-4 roadmap now shipped in 31c865c: Q&A modal,
selection search, tag filter, pi category filter, offline queue,
Related keyboard nav, Brain Ops dashboard, daily recall, MCP
cross-link
- Roadmap shortened to Phase-5 leftovers: jump-to-passage, wikilink
suggestions, memory explorer, MMR, multi-brain, canvas, conflict
detection
- New "Distribution" section pointing at ruvnet/obsidian-brain
(v0.1.0 tagged, release workflow auto-publishes assets)
README
- BRAT install now points at ruvnet/obsidian-brain instead of a path
into ruvnet/RuVector
Dist repo (ruvnet/obsidian-brain):
- Tag v0.1.0 pushed
- Release created with main.js (54.9 KB), manifest.json, styles.css,
versions.json
- Release + Test workflows pass green
Co-Authored-By: claude-flow <ruv@ruv.net>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implements ADR-SYS-0025. Ships an installable Obsidian plugin and an example vault under
examples/obsidian-brain/that bridges notes with the local RuVector brain (mcp-brain-server-localon127.0.0.1:9876) and embedder (127.0.0.1:9877).Features delivered
Cmd+Shift+Bsemantic searchsrc/search-modal.ts— DiskANN viaPOST /brain/search, fuzzy fallbacksrc/related-view.ts— refreshes on active-note changesrc/indexer.ts— debounced, content-hash deduped, AIDefence scannedsrc/bulk-sync.ts— include/exclude folders, progress barsrc/dpo.ts— mark chosen / rejected / export pairs tablesrc/graph-overlay.ts— color groups to.obsidian/graph.json, reversiblesrc/settings.ts,src/main.ts— live health + memory countExample / install
examples/obsidian-brain/example-vault/— minimal demo vaultexamples/obsidian-brain/systemd/{ruvector-brain,ruvector-embedder}.service— loopback-bound user unitsexamples/obsidian-brain/scripts/setup.sh <vault>— builds and dropsmain.js+manifest.json+styles.cssinto the target vaultTest plan
npm install && npm run buildproducesmain.js(30 KB) without TypeScript errorstsc --noEmit --skipLibCheckcleannode --check main.jspassesCmd+Shift+B, related panel, bulk-sync, DPO flow, graph overlay, status barBrain endpoints used
GET /health,GET /brain/info,GET /brain/index_statsPOST /brain/searchGET/POST /memories,GET /memories/:idPOST /security/scanGET/POST /preference_pairsPOST /embed(embedder service)Why an example, not a crate
The brain already ships DiskANN, AIDefence, embeddings and a stable HTTP surface. The plugin is intentionally thin — it's a reference client shared across CLI/MCP/agent rather than a parallel vector store.
🤖 Generated with claude-flow