chore: bump MCP package to v2.3.1, update changelog

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: Skiipy11

CHANGELOG.md

@@ -4,6 +4,20 @@ This changelog covers the entire Multi-Agent Memory project (API, MCP server, ad
 
 The canonical source is [`mcp-server/CHANGELOG.md`](mcp-server/CHANGELOG.md) -- this file mirrors it for convenience.
 
+## 2.3.1 (2026-03-29)
+
+### Bug Fixes
+- **brain_consolidate**: Fixed hardcoded `?sync=true` — async mode now works as documented. Added `sync` boolean parameter for explicit control.
+- **MCP input validation**: `brain_store`, `brain_search`, `brain_delete`, and `brain_client` now validate required arguments before calling the API, returning clear error messages instead of passing `undefined`.
+- **brain_export**: Added `limit` parameter (default 500) to prevent oversized responses. Description warns about large result sets.
+- **Server version**: Self-reported version now matches package.json.
+
+### Reliability
+- All API routes sanitize error messages (no internal details leaked to clients)
+- Fetch timeouts on all HTTP calls (embedders, LLM providers, Baserow, webhooks)
+- Export endpoint paginated, import endpoint now indexes keywords + extracts entities
+- Webhook dedup tenant-scoped, Postgres graceful shutdown, auth map cleanup
+
 ## 2.3.0 (2026-03-26)
 
 ### Multi-Path Retrieval with RRF Fusion

mcp-server/CHANGELOG.md

@@ -1,5 +1,19 @@
 # Changelog
 
+## 2.3.1 (2026-03-29)
+
+### Bug Fixes
+- **brain_consolidate**: Fixed hardcoded `?sync=true` — async mode now works as documented. Added `sync` boolean parameter for explicit control.
+- **MCP input validation**: `brain_store`, `brain_search`, `brain_delete`, and `brain_client` now validate required arguments before calling the API, returning clear error messages instead of passing `undefined`.
+- **brain_export**: Added `limit` parameter (default 500) to prevent oversized responses. Description warns about large result sets.
+- **Server version**: Self-reported version now matches package.json.
+
+### Reliability
+- All API routes sanitize error messages (no internal details leaked to clients)
+- Fetch timeouts on all HTTP calls (embedders, LLM providers, Baserow, webhooks)
+- Export endpoint paginated, import endpoint now indexes keywords + extracts entities
+- Webhook dedup tenant-scoped, Postgres graceful shutdown, auth map cleanup
+
 ## 2.3.0 (2026-03-26)
 
 ### Multi-Path Retrieval with RRF Fusion

mcp-server/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@zensystemai/multi-agent-memory-mcp",
-  "version": "2.3.0",
+  "version": "2.3.1",
   "type": "module",
   "description": "Persistent multi-path memory for AI agents — vector + BM25 keyword + entity graph search with RRF fusion, credential scrubbing, auto-consolidation, multi-backend storage (Qdrant + SQLite/Postgres)",
   "main": "src/index.js",

mcp-server/src/index.js

@@ -41,7 +41,7 @@ async function apiRequest(path, options = {}) {
 }
 
 const server = new Server(
-  { name: 'shared-brain', version: '2.3.0' },
+  { name: 'shared-brain', version: '2.3.1' },
   { capabilities: { tools: {} } }
 );