Skip to content

LIA

v1.21.24 Security

This release includes 2 security fixes for security teams reviewing exposed deployments.

✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →
This release patches 2 known CVEs

Topics

ai ai-assistant apple llm claude conversational-ai
+13 more
docker emotional-ai google healthcare-ai human-in-the-loop langchain langgraph mcp multi-agent personal-assistant self-hosted skills vision-ai

Affected surfaces

deps breaking_upgrade

Summary

AI summary

Updates GHSA-gr75-jv2w-4656, GHSA-g794-3fmp-753h, and https://github.com/jgouviergmail/LIA-Assistant/blob/main/CHANGELOG.md across a mixed release.

Full changelog

Security & code-scanning remediation

Full triage of the GitHub security surface — 76 code-scanning alerts, 24 Dependabot alerts, 15 dependency PRs — every finding traced to root cause instead of patching symptoms.

Security

  • Consolidated dependency bumps (24 alerts resolved): asyncssh 2.23.0 (GHSA-g794-3fmp-753h), langchain-anthropic 1.4.6 (GHSA-gr75-jv2w-4656), authlib 1.6.12, nltk 3.10.0, pygments 2.20.0 + manifest security floors (pyyaml, segno, python-pptx). Locks regenerated hash-verified, surgical diff. The 3 ecdsa alerts (Minerva, no upstream fix) dismissed with rationale: HS256-only JWT, cryptography backend, zero EC usage in src/.
  • Security workflow repaired: missing actions: read permission failed every CodeQL/Trivy upload since May 11, freezing the alert baseline; now also runs on every push to main.
  • Fresh-baseline follow-up: the first working scan surfaced findings accumulated during the blind window — all resolved: 8 CodeQL findings fixed at root cause (incl. RAG storage containment guard and the skills sandbox moving from world-writable 0777 to chown+0700), 11 frontend Trivy alerts fixed via pnpm overrides (dompurify 3.4.11, @grpc/grpc-js 1.14.4, protobufjs 7.6.5, @babel/core 7.29.7, vite 7.3.5).

Fixed

  • Assistant: "my next N appointments" no longer over-matches the daily-briefing skill (trigger-first description rewrite), verified live in pipeline and ReAct modes.
  • ReAct: the tool cap now keeps the detected domains' tools first and names every dropped tool — no more silent starvation, no more invented appointments.
  • Dependabot npm scope moved to the workspace root (frontend PRs can finally pass --frozen-lockfile).

Changed

  • All 193 except: pass handlers eliminated (CodeQL py/empty-except class closed) — contextlib.suppress doctrine + permanent AST guard in CI.
  • Node.js 22 → 24 LTS (web Docker images + CI).

Verified

8,921 backend unit tests green · MyPy strict · Docker boot healthy on the regenerated lock · 200/200 vitest on Node 24 · i18n parity ×6 languages · post-remediation scan: 0 open Dependabot alerts, 0 open code-scanning alerts (dismissals documented).

Full changelog: CHANGELOG.md

Docker Images

docker pull ghcr.io/jgouviergmail/LIA-Assistant/api:v1.21.24
docker pull ghcr.io/jgouviergmail/LIA-Assistant/web:v1.21.24

Quick Start

git clone https://github.com/jgouviergmail/LIA-Assistant.git
cd LIA-Assistant
cp .env.example .env
docker compose up -d

Full Changelog: https://github.com/jgouviergmail/LIA-Assistant/compare/v1.21.23...v1.21.24

🤖 Generated with Claude Code

Breaking Changes

  • Minimum runtime upgraded to Node.js 24 LTS (web Docker images + CI).

Security Fixes

  • GHSA-g794-3fmp-753h — asyncssh upgraded to 2.23.0 resolving vulnerability.
  • GHSA-gr75-jv2w-4656 — langchain-anthropic upgraded to 1.4.6 resolving vulnerability.

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

Track LIA

Get notified when new releases ship.

Sign up free

Beta — feedback welcome: [email protected]