# VerifyMCP

> Independent trust scores, tool surfaces and change history for MCP servers.

- name: io.verifymcp/mcp
- version: 1.0.0
- connection class: R0 (autonomous)
- trust: 92/100
- quality: 100/100 (strong)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/io.verifymcp%2Fmcp
- tools: https://api.protogrid.dev/v1/servers/io.verifymcp%2Fmcp/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "verifymcp": {
      "type": "http",
      "url": "https://mcp.verifymcp.io"
    }
  }
}
```


## Trust

- hygiene: 100
- liveness: 99
- freshness: 85
- provenance: 85
- drivers: +repository +dns-namespace +website +reachable +uptime-97% ~updated-46d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 100/100 (strong)

### protocol: 100 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- pass `protocol.list_ttl`: tool list is cacheable (ttlMs 3600000)

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 100 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- pass `tools.annotations`: 9 of 9 tools declare readOnlyHint or destructiveHint
- pass `tools.directory_hints`: every tool declares readOnlyHint, destructiveHint, idempotentHint and openWorldHint
- pass `tools.token_cost`: about 1,805 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: 3 tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 9, about 1,805 tokens to load them all
- tool set hash: fff41be21f68a39e3650b5225e8e6461
- tools last changed: 2026-10-16T22:40:39.781Z
- badge: [![protogrid quality](https://protogrid.dev/badge/io.verifymcp/mcp.svg)](https://protogrid.dev/servers/io.verifymcp/mcp)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-10-16 `get_server_alternatives` changed (description, outputSchema)
  - before: Other MCP servers doing a similar job to a given one, with their trust scores. Neighbours share a product or publisher namespace; this is not a semantic search.
  - after: Other MCP servers doing a similar job to a given one, with their trust scores. Neighbours share a product, publisher namespace or category; this is not a semantic search.
- 2026-09-27 `get_server_install_config` changed (description, outputSchema)
  - before: Ready-to-paste install snippets for one MCP server, for every client we support. Read get_server first: this returns configuration, not a safety judgement.
  - after: Ready-to-paste install snippets for one MCP server, for every client we support. An MCPB bundle has no launch command, so it returns the download URL, the registry's SHA-256 and commands to verify the file instead. Read get_server first: this returns configuration, not a safety judgement.
- 2026-09-24 `list_servers` changed (inputSchema)

Full history: https://api.protogrid.dev/v1/servers/io.verifymcp%2Fmcp/changes

## Remotes

- https://mcp.verifymcp.io (streamable-http, auth none, reachable true, uptime30d 0.9652174)

## Packages

_none_

## Tools (9)

- `get_server`: Get the full VerifyMCP trust report for one MCP server: score with per-category verdicts and reasons, the tools it exposes with their context-token cost, and its recent change history. Use before installing or trusting a server. Accepts a package name, endpoint URL, registry name or verifymcp slug.
- `get_server_alternatives`: Other MCP servers doing a similar job to a given one, with their trust scores. Neighbours share a product, publisher namespace or category; this is not a semantic search.
- `get_server_comparison`: Compare up to 5 MCP servers side by side: trust scores, per-category breakdown, tool counts and context-token cost. Use when choosing between candidates that do the same job. Accepts package names, endpoint URLs, registry names or verifymcp slugs.
- `get_server_diagnostics`: The evidence behind one server's score: TLS, DNSSEC, authorization, redirects, transports, provenance, install scripts, known CVEs and dependency health. Use to explain a low score.
- `get_server_install_config`: Ready-to-paste install snippets for one MCP server, for every client we support. An MCPB bundle has no launch command, so it returns the download URL, the registry's SHA-256 and commands to verify the file instead. Read get_server first: this returns configuration, not a safety judgement.
- `get_server_tools`: Every tool one MCP server exposes, with its parameters, output schema and context-token cost. Use when get_server reported the tool list was truncated.
- `list_known_malware`: List MCP server components carrying a supply-chain malware finding. Use to check whether anything in the register is flagged before installing.
- `list_products`: List the products MCP servers integrate with, such as github or notion, and how many servers cover each. Use to find the product slug that list_servers accepts.
- `list_servers`: Filter the VerifyMCP directory of scored MCP servers. Use to discover servers by name fragment, ecosystem, product or trust score. Name matching only: descriptions and capabilities are not searched, so a plain-English question matches nothing. Returns the first 100.