# Litmus

> Behavioral litmus harness for MCP servers — grade a server A–F (tool-output injection, egress, sensitive-data, adversarial-input) with reproducible, content-addressed evidence. Ships a CLI and an MCP server with a runlitmus tool for AI agents.

## Facts
- Page: https://tashan.sh/capability/pkg-polygraphso-litmus
- tashan id: pkg:@polygraphso/litmus
- Source: https://github.com/polygraphso/litmus
- npm: https://www.npmjs.com/package/@polygraphso/litmus
- Type: npm
- Category: security
- tashan score: 48.0 / 100
- Adoption: 25.0
- Upkeep: 67.0
- Freshness: 82.0
- Evidence coverage: 100% of the inputs this score can use
- Health: active
- Instruction depth: not yet graded
- npm downloads: 64/week
- License: Apache-2.0
- Official: no

## Install

```sh
claude mcp add polygraphso-litmus -- npx -y @polygraphso/litmus
```

## Security audit
- Known advisories: 0
- Install-time script: none declared
- Build provenance: not attested

Permissions are read from DECLARED dependencies only. Nothing is executed, so an empty result means "nothing declared", never "nothing possible".

---
Measured 2026-09-12 by tashan (https://tashan.sh) from public evidence. Scorer s5.