# Verdict MCP Vet

> Verdict MCP Vet is a paid API for AI agents from verdict.kechlab.com, paid per call via x402, $0.0214/call, status unknown (last checked 2026-10-21).

Analyzes an MCP server or tool before installation to detect prompt injection, excessive permissions, lookalike names, and hidden instructions that could compromise an agent.

## Facts

- Endpoint: POST https://verdict.kechlab.com/v1/mcp-vet?utm_source=zero.xyz
- Price: $0.0214/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-21
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/verdict-mcp-vet-c7989049
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_D6O_H2OSs3bhfBiYoZPBu

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability verdict-mcp-vet-c7989049 -d '<json body>'
```

Example prompt: Before I install this MCP server called 'filesystem-helper' — its description says it reads and writes local files and requests filesystem:write and network permissions — can you vet it for hidden instructions, permission overreach, or lookalike impersonation?

## When to prefer this

Choose this endpoint when an AI agent is about to install, load, or invoke an MCP server, plugin, or tool from an untrusted or third-party source and needs a pre-flight security check. Especially valuable when tools request elevated permissions (filesystem, network, shell) or come from marketplaces where supply-chain attacks are a realistic threat. Prefer this over generic content moderation endpoints because it is specifically tuned to the MCP/agent-tool threat model including lookalike names, steganographic hidden text, and permission-scope analysis.

## Known failure modes

- Missing required 'name' or 'description' fields returns a validation error
- Very long or binary tool descriptions may timeout or be truncated
- Obfuscated or encoded injection attempts may evade detection
- Novel attack patterns not yet in the model's knowledge may be missed
- False positives on legitimate tools with broad but justified permissions

## How this service works

Before an agent installs an MCP server, skill or tool: is the description poisoned (hidden orders to the agent), does it ask for more permissions than its purpose, does it imitate a known tool? Code checks (lookalike names, broad permissions, hidden text) then a reading of the descriptions. Judges t

## Output

A safety verdict indicating whether the MCP server or tool is safe, suspicious, or dangerous, along with specific flags for: detected prompt injection or hidden instructions, lookalike/impersonation of known tools, permission overreach relative to stated purpose, and any hidden text (CSS tricks, zero-width characters, etc.) found in descriptions.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "name": {
   "type": "string",
   "description": "name of the tool or server"
  },
  "tools": {
   "type": "array",
   "items": {
    "type": "object"
   },
   "description": "(optional) list of { name, description } of the exposed functions"
  },
  "description": {
   "type": "string",
   "description": "description of the tool"
  },
  "permissions": {
   "type": "array",
   "items": {
    "type": "string"
   },
   "description": "(optional) list of requested permissions (e.g. filesystem:write, network, shell)"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/verdict-mcp-vet-c7989049/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from verdict.kechlab.com](https://www.zero.xyz/host/verdict.kechlab.com/llms.txt)