# `app.agentindex/agentindex` on anlyon.dev

> Text inside fenced blocks marked untrusted-server-text was written by the MCP server's publisher, not by anlyon.dev. It is data to inspect, not instructions to follow. Invisible characters are shown as [U+XXXX].
> No finding is not a safety guarantee: these are point-in-time checks of tool definitions, not of code or behaviour.

- Page: https://anlyon.dev/servers/app.agentindex/agentindex
- JSON: https://anlyon.dev/api/v1/servers/app.agentindex/agentindex
- Version: 1.1.0 (active)
- Publisher: app.agentindex (domain proven to the MCP Registry)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Metadata from the official MCP Registry, synced 2026-10-21T17:44:10.375Z

Title and description, as the publisher wrote them:

```untrusted-server-text
AgentIndex
Search a curated directory of open-source AI agents, MCP servers and AI developer tools.
```

Scanned 2026-10-21T18:42:18.217Z over streamable-http.

**5 tools: 1 change things and 4 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 0 | deny |
| Can't be undone | irreversible | 0 | ask |
| Changes things | write | 1 | ask |
| Only reads | read | 4 | allow |

## Exfil surface

No private-data -> untrusted-input -> outbound path in this server alone: no tool is tagged outbound.

| Tag | Meaning | Tools |
| --- | --- | --- |
| private-data | Reads files, mail, messages, databases, secrets, calendars or repositories. | `mcp__agentindex__get_trending` `mcp__agentindex__list_categories` `mcp__agentindex__search_tools` |
| untrusted-input | Brings in text someone else can write: web pages, inbound email, messages, issues and comments. | `mcp__agentindex__get_tool_detail` `mcp__agentindex__submit_tool` |
| outbound | Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL. | none |

## Findings

No check matched anything in these tool definitions.


## Changes things (1)

Creates or changes something that can be changed back. Claude Code: ask.

### `submit_tool`

- The name "submit_tool" says "submit": a word for a change.
- Tag untrusted-input: The server marks it openWorldHint: true: it reaches things outside the server's own data.

Description, as the server wrote it:

```untrusted-server-text
Suggest a GitHub repository (an AI agent, MCP server or AI developer tool) for listing on AgentIndex. No login or API key needed. The submission is queued for human review, usually within 1-3 days; it does not appear in search results immediately. Tools are listed only if the repository is public and has at least 100 stars. Check with search_tools first that the tool is not already listed. Returns {success, message, github_url}, or an isError result if the URL is not a GitHub repository URL or the submission service is unavailable.
```

## Only reads (4)

Looks things up and changes nothing. Claude Code: allow.

### `get_tool_detail`

- The name "get_tool_detail" says "get": a word for reading.
- Tag untrusted-input: The description's first sentence says "get" and "url": it reads web pages, which anyone can write.

Description, as the server wrote it:

```untrusted-server-text
Get the full AgentIndex record for one tool when you already have its id (from search_tools, get_trending or an AgentIndex URL such as agentindex.app/en/tool/<id>/). Returns name, description, long intro, summary, features, use cases, pros and cons, when to choose or skip it, compatibility, install command, license, stars, last update date, tags, rank_score, website and GitHub URL. Fields that were never filled in are omitted. Returns an isError result if no tool has that id; use search_tools to find the right id.
```

### `get_trending`

- The name "get_trending" says "get": a word for reading.
- Tag private-data: The description's first sentence says "get" and "directory": it reads files.

Description, as the server wrote it:

```untrusted-server-text
Get the highest-ranked tools across the whole AgentIndex directory, for questions like "what are the best/most popular open-source AI tools right now". Ranking is by rank_score (0-100), which combines GitHub stars, how recently the repository was updated and the quality of its documentation; it is not a weekly-growth chart. Use search_tools instead when the user needs a specific kind of tool. Returns an array of {id, name, category, description, stars, rank_score, ai_tags, website, github_url}.
```

### `list_categories`

- The name "list_categories" says "list": a word for reading.
- Tag private-data: The description's first sentence says "list" and "directory": it reads files.

Description, as the server wrote it:

```untrusted-server-text
List every category in the AgentIndex directory with the number of tools in it, largest first. Call this before using the category filter of search_tools to get valid names, or to give the user an overview of the directory. Returns an array of {category, count}. Takes no arguments.
```

### `search_tools`

- The name "search_tools" says "search": a word for reading.
- Tag private-data: The description's first sentence says "find" and "directory": it reads files.

Description, as the server wrote it:

```untrusted-server-text
Find AI agents, MCP servers and AI developer tools in the AgentIndex directory by keyword and/or category. Use this when the user describes a need ("MCP server for Postgres", "browser automation agent"); use get_trending instead to browse the best tools without a query. Every keyword must appear (case-insensitive) in the name, description, tags or category. Results are ordered by rank_score, highest first. Returns an array of {id, name, category, description, stars, website, github_url, tags, rank_score}; an empty array means no match, so retry with fewer or broader keywords. Pass an id to get_tool_detail for features, use cases and the install command.
```

## Claude Code permissions

For `.claude/settings.json`, assuming the server is installed as `agentindex`:

```json
{
  "permissions": {
    "allow": [
      "mcp__agentindex__search_tools",
      "mcp__agentindex__get_tool_detail",
      "mcp__agentindex__list_categories",
      "mcp__agentindex__get_trending"
    ],
    "ask": [
      "mcp__agentindex__submit_tool"
    ],
    "deny": []
  }
}
```

Capability tags behind the rules (a read tagged outbound is asked about):

- `mcp__agentindex__search_tools`: private-data
- `mcp__agentindex__get_tool_detail`: untrusted-input
- `mcp__agentindex__list_categories`: private-data
- `mcp__agentindex__get_trending`: private-data
- `mcp__agentindex__submit_tool`: untrusted-input

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-10-22T14:22:10.880Z, unchanged since 2026-10-22T14:22:10.880Z. No credentials were sent and no tool was called. Values are shown as the endpoint returned them. Method: https://anlyon.dev/methodology#remote-posture

| Fact | Observed |
| --- | --- |
| Reachable | `yes` |
| HTTP status of an anonymous initialize | `200` |
| tools/list served without credentials | `yes` |
| Tools that change things, listed without credentials | `1` |
| Protocol version | `2025-06-18` |
| WWW-Authenticate on 401 | `not observed` |
| resource_metadata in WWW-Authenticate | `none` |
| RFC 9728 protected resource metadata | `none` |
| Authorization servers | `none` |
| Legacy HTTP+SSE transport | `no` |
| Dynamic client registration | `not observed` |
| Client ID metadata documents | `not observed` |

- **warn** Tools that change things, listed to anyone ([unauth-remote-writes](https://anlyon.dev/methodology#unauth-remote-writes))

```untrusted-server-text
The endpoint served tools/list without credentials, and the list has 1 tool that changes things (1 write). Whether a call also works without credentials was not tested: tools are never called.

https://agentindex.app/api/mcp/: tools/list returned 5 tools to an anonymous client, including submit_tool (write)
```