This release includes 1 security fix for security teams reviewing exposed deployments.
Topics
+8 more
Affected surfaces
Summary
AI summaryFix SQL injection vulnerability in vector table creation.
Full changelog
v1.8.0 (2026-03-17)
This release is published under the MIT License.
Bug Fixes
-
Add missing error test in server embedding initialization (#203,
0eb3ef5) -
Add missing OSError test for chmod in token store (#199,
8d52eed) -
Disable mise runtime updates in Renovate (
7f99704) -
Fix Potential Command Injection in subprocess.run (#204,
c1a8012) -
Fix SQL injection vulnerability in vector table creation (#208,
7a95b37) -
Remove mcp-name entry from README (
dfe85d7) -
ci: Use pull_request_target for jobs requiring secrets (
823944a) -
deps: Update dependency qwen3-embed to >=1.4.3 (#188,
bad4d15)
Chores
-
Add glama.json for Glama directory listing (
6bdc37d) -
Standardize repo files across MCP server portfolio (
ce4a6b7) -
deps: Update actions/download-artifact digest to 3e5f45b (#187,
32e0d32) -
deps: Update astral-sh/setup-uv digest to 37802ad (#194,
2c0b989) -
deps: Update dawidd6/action-send-mail action to v15 (#209,
ce2b5a5)
Code Style
Documentation
- Add v1.8-v1.9 design spec (
b19ed64)
Features
-
Add better-telegram-mcp to Also by section and mcp-name (
5aa65f0) -
Add Glama.ai badge to README (
fb9981f) -
Add Jina AI embedding priority and dual-backend reranker (
02f02cb) -
Add knowledge graph, importance scoring, archive, and dedup (
57c44fb) -
Offload blocking SQLite I/O to thread in sync_full (#202,
e34c2f8) -
Testing improvement] Add test for invalid JSON token in setup_sync (#200,
e683697) -
Wire intelligence features into server (graph, importance, archive, consolidate) (
351475c)
Detailed Changes: v1.7.0...v1.8.0
Security Fixes
- Fix SQL injection vulnerability in vector table creation
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About n24q02m/mnemo-mcp
Persistent AI memory with SQLite hybrid search (FTS5 + semantic). Built-in Qwen3 embedding, rclone sync across machines. Zero config, no cloud, no limits.
Related context
Beta — feedback welcome: [email protected]