# ankitkapur1992-hlido/hlido-mcp [Health: Active]

**Category:** 🔬 Research  
**Repository:** https://github.com/ankitkapur1992-hlido/hlido-mcp  
**GitHub Stars:** 0  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/ankitkapur1992-hlido-hlido-mcp

## Description
Independent trust scores, claim audits, and comparisons for AI agents — queryable by your agent over MCP. Hosted Cloudflare Worker at hlido.eu/mcp (no install). Returns a 0–100 score, tier verdict, per-claim PASS/FAIL audit, and signed evidence for a reviewed agent. From Hlido.

## Tools
Capabilities this server exposes over MCP:

- **trust_check** — "Is agent X trustworthy?" — score, tier, verdict for a slug
- **find_trusted** — "Find me a trusted agent for <need>" — filtered registry search
- **verify_claim** — "Does X really do Y?" — per-claim PASS/FAIL evidence
- **compare_agents** — Side-by-side scorecard comparison
- **get_scorecard** — Full sanitized scorecard JSON for a slug
- **find_similar_agents** — Semantic nearest neighbours to a given agent
- **submit_agent** — Nominate an agent for review
- **report_review_issue** — Flag a problem with a published review
- **request_quick_audit** — Ask for a fast re-check of a stale review

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "hlido-mcp": {
    "command": "npx",
    "args": ["-y","wrangler"]
  }
}
```

## Documentation

## What ankitkapur1992-hlido/hlido-mcp MCP server does

The ankitkapur1992-hlido/hlido-mcp MCP server connects MCP clients to Hlido’s public registry of reviewed AI agents. It is intended for agents or applications that need to assess whether another agent is suitable for a task, inspect a published review, or compare several candidates.

A review can provide a numeric score from 0 to 100, a tier and verdict, claim-level audit results, and public supporting evidence. The server exposes sanitized scorecard information rather than the private scoring methodology, grader assertions, or editorial drafts.

## How it works

The hosted service is available at `https://hlido.eu/mcp` and communicates through MCP over HTTP. It reads the public JSON published by Hlido, including registry, scorecard, and attestation data. Requests are anonymous and do not require credentials. Lightweight daily counters are used for per-tool telemetry, according to the project documentation.

The server is an adapter over published review data. The review pipeline, agent testing engine, and scoring model are outside this repository, so using the MCP server does not provide access to those internal systems.

## Setup and configuration

The simplest setup uses the hosted endpoint and requires no local installation. Add `https://hlido.eu/mcp` as an HTTP MCP server in a compatible client. The README provides configuration examples for Claude Desktop, Cursor, and other MCP clients, as well as a Claude Code command.

The ankitkapur1992-hlido/hlido-mcp MCP server can also be self-hosted as a Cloudflare Worker. A self-hosting setup copies `wrangler.toml.example` to `wrangler.toml`, adds the Cloudflare account ID, and deploys with Wrangler. Optional KV cache and Vectorize similarity-index bindings can be omitted; the worker then fetches public JSON directly.

For local testing, the repository includes a Docker workflow. Building the image and running it with port 8080 exposed starts the worker on the local `workerd` runtime through Wrangler development mode. The local MCP endpoint is `http://localhost:8080/`.

## Tools and capabilities

The ankitkapur1992-hlido/hlido-mcp MCP server exposes these documented operations:

- `trust_check`: retrieve a score, tier, and verdict for an agent slug.
- `find_trusted`: search the registry for agents matching a need.
- `verify_claim`: check whether a reviewed agent supports a stated claim and return PASS/FAIL evidence.
- `compare_agents`: view side-by-side scorecards.
- `get_scorecard`: retrieve the complete sanitized scorecard JSON for a slug.
- `find_similar_agents`: find semantic nearest neighbours for an agent.
- `submit_agent`: nominate an agent for review.
- `report_review_issue`: flag a problem with an existing review.
- `request_quick_audit`: request a fast re-check of a stale review.

The repository also mentions additional discovery and metadata tools in its live tool table.

## Limitations and notes

Results are limited to information published in Hlido’s public data sources. The server does not reveal scoring weights, private grader assertions, or editorial drafts. It also does not perform the underlying agent reviews or testing itself.

The hosted endpoint is the recommended no-install path. Self-hosting requires a Cloudflare Worker setup, while the Docker option runs the worker locally without requiring a Cloudflare account. The code is licensed under Apache-2.0; the review data is identified separately as CC-BY.

_Full upstream README: https://allmcps.com/mcp/ankitkapur1992-hlido-hlido-mcp/readme_