Latex Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Latex Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
An open-source Model Context Protocol server that lets LLMs (Claude, Codex, Gemini/Antigravity, and other MCP-compatible clients) compile LaTeX documents and inspect the resulting PDF — enabling a fully autonomous write → compile → check → revise loop.
It talks to a self-hosted instance of Overleaf's open-source CLSI (Common LaTeX Service Interface) over HTTP, and also exposes a small web UI (in the spirit of Overleaf's editor) so a human can watch or collaborate on the same document a connected LLM is editing.
Status: early scaffolding, under active development.
[LLM client (Claude / Codex / Antigravity / ...)] [Browser UI]
| MCP (stdio) | HTTP + WebSocket
v v
[latex-mcp backend service]
|
| HTTP REST
v
[Self-hosted CLSI] (Docker)
|
v
[TeX Live container] (actual compilation)Full setup instructions (CLSI via Docker, running the backend, and connecting from Claude Desktop, Claude Code, Codex CLI, Antigravity, and others) are coming as the project is built out.
MIT — see LICENSE. Note this project is independent and not affiliated with Overleaf; it only communicates with the open-source CLSI component over its HTTP API.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.