What is overleaf-mcp CLAUDE.md?

A CLAUDE.md instruction file explaining how the Overleaf MCP server edits LaTeX project files through Overleaf’s Git connection.

Use it when maintaining the server or its write tools. It documents pulling changes, checking paths, editing files, committing, and pushing updates to Overleaf.

It clarifies the supported write process and prevents contributors from looking for an Overleaf REST API that the project does not use.

How amcheste/overleaf-mcp configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to amcheste/overleaf-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/amcheste/overleaf-mcp/develop/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/amcheste/overleaf-mcp

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/amcheste/overleaf-mcp/claude-md/github.svg)](https://agentmods.dev/instructions/amcheste/overleaf-mcp/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/amcheste/overleaf-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/amcheste/overleaf-mcp/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/amcheste/overleaf-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/amcheste/overleaf-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 492 This file is loaded in full into every session.

When invoked 492 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00492	$0.00492
Opus 5.5	$0.00197	$0.00197
Sonnet 5.5	$0.00098	$0.00098
Haiku 4.5	$0.00049	$0.00049

Measured 2mo ago against content hash 6354718085f1, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

overleaf-mcp CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 39 lines

How it starts

The opening of the file, as written. The whole thing — 39 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Overleaf MCP Server: architecture notes

Project-local context for Claude sessions. The README.md is for end users; this file is for contributors (human or AI) working on the server itself.

The write path IS git

Overleaf does not expose a public HTTP write API for project file trees. The only supported programmatic write path on a paid Overleaf account is the per-project Git remote at git.overleaf.com/<project_id>.

This server is therefore a local git client:

Claude ──MCP─► overleaf-mcp ──git-over-HTTPS──► git.overleaf.com
               (cached local clone)

Every write tool (edit_file, create_file, delete_file) follows the same flow:

git pull to refresh the cached clone
Validate the target path stays inside the repo root
Write/delete the file in the working tree
git commit with a generated message and the configured author
git push to Overleaf

If you ever find yourself reaching for an "Overleaf REST API" call, stop. There isn't one. The answer is always a git operation against the cached clone.

We shell out to git via subprocess rather than using GitPython. The design goal is auditability: the dependency graph should be readable in five minutes.

GitHub mirroring is out of scope

Users who want a GitHub backup add GitHub as a second remote on their own local clone. The MCP server knows nothing about GitHub and should not grow GitHub-aware code. If this boundary ever gets tempted, push back.

Credentials never touch disk

Tokens live in the OS keychain (primary) or environment variables (fallback, for CI/power users). The config file contains project aliases + IDs only, never tokens. See src/overleaf_mcp/core/credentials.py for the resolution order.

Scope discipline

This server is deliberately single-process, single-user, stdio-transport, serial. No Redis, no job queues, no locks, no concurrent-edit handling. Those assumptions keep the code auditable and are why we can ship something meaningfully different from the existing community servers. If you find yourself adding coordination machinery, stop and re-read this section.

Read the full file on GitHub · 39 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2mo ago First seen · 39 lines · 492 tokens per session scan A 6354718085f1

Subscribe to this mod's changes

overleaf-mcp CLAUDE.md is an instructions file published in the GitHub repository amcheste/overleaf-mcp (0 stars, last pushed 28d ago), licensed MIT. It adds 492 tokens to every session, about $0.0020 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

LaTeX academic-writing claude claude-desktop mcp mcp-server

overleaf-mcp CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
overleaf-mcp CLAUDE.mdamcheste/overleaf-mcp	Instructions file	0	28d ago	MIT	A	492	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

overleaf-mcp: Instructions file for Claude Code