Skip to content

armgit / ruflo

Public Imported from GitHub
0

🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated

fix(memory): reject 128-dim bridge results — MCP no longer emits mock embeddings (#2395)

ruv04f47445068d
diff --git a/v3/@claude-flow/cli/src/memory/memory-bridge.ts b/v3/@claude-flow/cli/src/memory/memory-bridge.ts
index 3fce82ba11..694cee674d 100644
--- a/v3/@claude-flow/cli/src/memory/memory-bridge.ts
+++ b/v3/@claude-flow/cli/src/memory/memory-bridge.ts
@@ -1191,6 +1191,19 @@ export async function bridgeGenerateEmbedding(
const isMock = (embedder as { isMock?: boolean; backend?: string }).isMock === true
|| (embedder as { backend?: string }).backend === 'mock';
+ // #2395 — AgentDB's vectorBackend controller may initialize as
+ // `enabled: false` (e.g., ruvector isn't wired into its neural substrate
+ // yet) and silently fall back to a 128-dim hash stub. The embedder doesn't
+ // expose isMock=true in that path, so the wrapper above was reporting
+ // backend='onnx' on what's actually 128-dim hash garbage. The model name
+ // we hardcode ('Xenova/all-MiniLM-L6-v2') always produces 384-dim, so a
+ // dimensions mismatch is a definitive stub signal: return null and let
+ // the caller fall through to generateLocalEmbedding which routes to
+ // transformers.js / ruvector ONNX directly.
+ if (emb.length !== 384) {
+ return null;
+ }
+
return {
embedding: Array.from(emb),
dimensions: emb.length,