arxivsubSAFE
arxivsub-mcp helps you find papers on arXivSub effortlessly
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
An MCP server for [arXivSub](https://arxivsub.comfyai.app) — search arXiv and the top AI / CV / ML / robotics / NLP conferences, follow any paper to its related work, build research analytics, and manage a personal library, straight from your AI agent (Claude Desktop, Cursor, Cline, Claude Code, Windsurf, ...).
arXivSub summarizes every paper into a structured 6-part breakdown (what it's about, innovations, techniques, datasets, results, limitations) and indexes it for semantic search across arXiv + CVPR, ICCV, ECCV, ICLR, ICML, NeurIPS, AAAI, MICCAI, CoRL, RSS, ACL, EMNLP, ICRA, IROS, IJCAI, SIGGRAPH, SIGGRAPH Asia. This server puts that engine directly in your agent's hands.
Get a free key (7-day trial, no credit card)
- Sign in at https://arxivsub.comfyai.app/skills with email or Google.
- Copy your API key from the Skills page.
- Put it in the server config below as
ARXIVSUB_SKILL_KEY.
New accounts get a 7-day free trial of the full skill. After that it's part of arXivSub Pro.
Install
The server runs via npx — no manual install needed. Add it to your MCP client's config.
Claude Desktop (claude_desktop_config.json), Cursor (~/.cursor/mcp.json), Windsurf, Cline, etc.:
{
"mcpServers": {
"arxivsub": {
"command": "npx",
"args": ["-y", "arxivsub-mcp"],
"env": { "ARXIVSUB_SKILL_KEY": "your_key_here" }
}
}
}Claude Code:
claude mcp add arxivsub --env ARXIVSUB_SKILL_KEY=your_key_here -- npx -y arxivsub-mcp
Codex CLI:
codex mcp add arxivsub --env ARXIVSUB_SKILL_KEY=your_key_here -- npx -y arxivsub-mcp
Or add it to ~/.codex/config.toml directly:
[mcp_servers.arxivsub]
command = "npx"
args = ["-y", "arxivsub-mcp"]
env = { ARXIVSUB_SKILL_KEY = "your_key_here" }Restart your client and the tools below become available.
claude.ai (remote connector)
On [claude.ai](https://c
f16ef5000aa7OBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add arxivsub-mcp --env ARXIVSUB_SKILL_KEY=${ARXIVSUB_SKILL_KEY} -- npx -y [email protected]Exposed tools (5)
4 read · 1 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
find_similar_papers | read | Given ONE paper you already have (from a prior search result), return the most similar papers |
library_list | read | List the user |
library_save | write | Save (or update) a paper in the user |
research_insight | read | Build aggregate pivot statistics over the arXivSub corpora — the same engine as the website |
search_papers | read | Search academic papers from arXiv and major AI/CV/ML/robotics/NLP conferences via arXivSub |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (2 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (2)
@modelcontextprotocol/sdk, zod, @types/node, typescript
- **Privacy / security:** the server only holds your skill key and talks to the arXivSub gateway over HTTPS. It has no direct database access and no service credentials.
Gates applied: no_behavioural_pass.
f16ef5000aa7full audit observations/trust-audit/mcp-server/arxivsub__arxivsub-mcp.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | f16ef5000aa7 | SAFE | B | 89 | first audit |
Questions
What is the arxivsub MCP server?
arxivsub-mcp helps you find papers on arXivSub effortlessly
What tools does arxivsub expose?
5 in total: 4 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is arxivsub safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does arxivsub need?
It reads ARXIVSUB_SKILL_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does arxivsub run?
It speaks stdio and streamable-http, so it runs as a local process your client starts. It is published on npm as arxivsub-mcp at 0.1.1.
How current is this page?
The grade is for one exact copy of the source (f16ef5000aa7), read on 2026-10-12. The repository is watched and re-audited when it changes.