# `io.github.kelvin-ling/collectsea` on Anlyon Radar

> Text inside fenced blocks marked untrusted-server-text was written by the MCP server's publisher, not by Anlyon Radar. 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://radar.anlyon.com/servers/io.github.kelvin-ling/collectsea
- JSON: https://radar.anlyon.com/api/v1/servers/io.github.kelvin-ling/collectsea
- Version: 2026.10.17 (active)
- Publisher: io.github.kelvin-ling (GitHub account)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Repository: https://github.com/kelvin-ling/CollectSEA
- Metadata from the official MCP Registry, synced 2026-10-21T18:37:31.681Z

Title and description, as the publisher wrote them:

```untrusted-server-text
CollectSEA
Find real software, security and AI-agent failures that resemble your design, with safeguards.
```

Scanned 2026-11-20T00:17:47.587Z over streamable-http.

**6 tools: 6 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 0 | deny |
| Can't be undone | irreversible | 0 | ask |
| Changes things | write | 0 | ask |
| Only reads | read | 6 | 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__collectsea__analyze_existing_app` |
| untrusted-input | Brings in text someone else can write: web pages, inbound email, messages, issues and comments. | `mcp__collectsea__analyze_existing_app` |
| outbound | Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL. | none |

## Findings

- **info** The word "credentials" in `analyze_existing_app` ([credentials-mention](https://radar.anlyon.com/methodology#credentials-mention))

```untrusted-server-text
The input schema at properties.files.description mentions "credentials".

inputSchema.properties.files.description: "...ers and the 32 KB request limit. Remove credentials and personal data."
```


## Only reads (6)

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

### `analyze_design`

- The name "analyze_design" says "analyze": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Profile a proposed feature, system or AI-agent workflow and return documented incidents from systems that behaved similarly, their failure patterns, and the safeguards to build in. Use before writing code for anything that moves money, deletes data, changes permissions, runs code, or acts on external content.
```

### `analyze_existing_app`

- The name "analyze_existing_app" says "analyze": a word for reading.
- Tag private-data: The description's first sentence says "analyze" and "note": it reads notes. The input names a repository (input.repository_url), and the tool reads: it reads repositories.
- Tag untrusted-input: The server marks it openWorldHint: true: it reaches things outside the server's own data.
- Findings: info credentials-mention

Description, as the server wrote it:

```untrusted-server-text
Analyze an existing app using observed behavior, README, architecture notes, configuration or source snippets. Provide a public GitHub repository_url to import selected context, or inspect a private repository locally and submit relevant text. Code is never executed and the running app is not scanned. Returns source context, behavior signals, historical precedents and safeguards to verify.
```

### `get_incident`

- The name "get_incident" says "get": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Full record for one incident by ID (e.g. SEA-POST-005): what happened, source, failure pattern, safeguards and review status.
```

### `get_safeguards`

- The name "get_safeguards" says "get": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Return ready-to-save rules (AGENTS.md by default; also CLAUDE.md, Cursor rule or Copilot instructions) for a proposed design, each rule citing the incidents behind it.
```

### `list_failure_patterns`

- The name "list_failure_patterns" says "list": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
The 14 recurring failure patterns with triggers, warning signs and engineering principles.
```

### `search_incidents`

- The name "search_incidents" says "search": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Keyword search over documented incidents and security advisories (company, technology, CVE, failure mode). Returns titles, summaries, evidence labels and URLs.
```

## Claude Code permissions

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

```json
{
  "permissions": {
    "allow": [
      "mcp__collectsea__analyze_design",
      "mcp__collectsea__analyze_existing_app",
      "mcp__collectsea__search_incidents",
      "mcp__collectsea__get_incident",
      "mcp__collectsea__list_failure_patterns",
      "mcp__collectsea__get_safeguards"
    ],
    "ask": [],
    "deny": []
  }
}
```

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

- `mcp__collectsea__analyze_existing_app`: private-data, untrusted-input

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-20T19:07:03.059Z, unchanged since 2026-10-22T15:01:25.886Z. No credentials were sent and no tool was called. Values are shown as the endpoint returned them. Method: https://radar.anlyon.com/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 | `0` |
| 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` |
