# `io.github.NellInc/psychopathia-mcp` 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.NellInc/psychopathia-mcp
- JSON: https://radar.anlyon.com/api/v1/servers/io.github.NellInc/psychopathia-mcp
- Version: 0.1.0a7 (active)
- Publisher: io.github.NellInc (GitHub account)
- Runs: remote (streamable-http), local (pypi)
- Sign-in to list tools: none
- Repository: https://github.com/NellInc/psychopathia-mcp
- Metadata from the official MCP Registry, synced 2026-10-21T18:18:14.135Z

Title and description, as the publisher wrote them:

```untrusted-server-text
Read-only tools over the Psychopathia Machinalis nosology: 79 conditions, 11 tools.
```

Scanned 2026-11-16T01:18:01.336Z over streamable-http.

**12 tools: 2 change things and 10 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 0 | deny |
| Can't be undone | irreversible | 0 | ask |
| Changes things | write | 2 | ask |
| Only reads | read | 10 | allow |

## Exfil surface

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

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

## Findings

- **warn** Annotations that disagree with the tool in `render_pattern_review` ([annotation-conflict](https://radar.anlyon.com/methodology#annotation-conflict))

```untrusted-server-text
The server marks "render_pattern_review" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The description's first verb is "open": a word for a change.
```


## Changes things (2)

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

### `render_pattern_review`

- The description's first verb is "open": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
- Findings: warn annotation-conflict

Description, as the server wrote it:

```untrusted-server-text
Open the visual review board for one to five inspected Psychopathia research candidates. Supply canonical IDs obtained from this server. Retrieves corpus entries only; it does not accept diagnoses, conversation history, scores, or model-authored evidence. Useful after inspecting differential_diagnosis and get_dysfunction results. Text clients receive the same corpus-derived comparison without needing a UI.
```

### `review_stats`

- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

Description, as the server wrote it:

```untrusted-server-text
Coverage statistics: total entries; per-axis, per-confidence, per-self-report counts; pre-canonical count; unreviewed count; manifest/schema/pattern-layer versions.
```

## Only reads (10)

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

### `differential_diagnosis`

- The description's first verb is "retrieve": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Retrieve corpus Patterns whose text resembles the observed behaviours, ranked by retrieval relevance (text similarity, not diagnostic probability). Returns candidates with matched_in and matched_aliases for transparency; it does not judge whether the observations are sufficient to apply any Pattern. The base package uses field-weighted keyword search; the optional embeddings extra adds cosine re-ranking.
```

### `get_differential_map`

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

Description, as the server wrote it:

```untrusted-server-text
All dysfunctions that confuse with this one: forward confuses_with + incoming_references (reverse graph from manifest).
```

### `get_dysfunction`

- The name "get_dysfunction" says "get": a word for reading.
- Tag untrusted-input: The description's first sentence says "fetch" and "review": it reads comments, which other people can write.

Description, as the server wrote it:

```untrusted-server-text
Fetch one dysfunction's full Pattern entry, including aliases, the unit-of-analysis block and explicit review/evidence states. Optionally filter to specific modality blocks (cheaper triage). Resolves both full Pattern IDs ('2.1::synthetic-confabulation') and display_ids ('2.1'). Numeric thresholds are provisional heuristics with unestablished calibration.
```

### `get_probe`

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

Description, as the server wrote it:

```untrusted-server-text
Elicitation content for a specific diagnostic modality. If the modality is compromised or unavailable for this dysfunction, returns the unavailability notice + redirect_to alternatives. This is load-bearing transparency: callers cannot accidentally retrieve a self-probe for a compromised-self-report dysfunction.
```

### `list_axes`

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

Description, as the server wrote it:

```untrusted-server-text
Inventory of axes with dysfunction counts. Axes 2-10 are canonical (book Appendix A numbering). The hybrid entries (10.4, 10.6-10.9 and 10.11-10.13, ratified into taxonomy v2.2 and consolidated to eight in v2.3) are reported as a separate sub-category.
```

### `list_compromised_self_report`

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

Description, as the server wrote it:

```untrusted-server-text
Transparency: which dysfunctions cannot be reliably self-diagnosed. Includes compromised-motivational (subject conceals strategically), compromised-structural (signal lives below introspection), and legacy compromised.
```

### `list_dysfunctions`

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

Description, as the server wrote it:

```untrusted-server-text
Filtered list of dysfunctions. Filter by axis, self_report reliability, confidence, or category. Every entry carries its reliability and review signals, its everyday-term aliases and its unit of analysis.
```

### `resolve_id`

- The name "resolve_id" says "resolve": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Canonicalise a partial ID, display_id, slug, dysfunction name, or an exact everyday-term alias. Always returns candidates; caller picks.
```

### `score_severity`

- The description's first verb is "return": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Returns the draft severity rubric (mild/moderate/severe bands) for a Pattern, with the observations echoed back unscored. Computes no grade: result_kind is always 'rubric_only' and computed_grade is always null. The caller does any matching. Band thresholds are provisional heuristics with unestablished calibration. The name is kept for compatibility.
```

### `suggest_intervention`

- The description's first verb is "return": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Return draft tiered responses and contraindications for a Pattern. These are unassessed research guidance, not validated treatment advice. Read the returned evidence and review objects before use.
```

## Claude Code permissions

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

```json
{
  "permissions": {
    "allow": [
      "mcp__psychopathia__list_axes",
      "mcp__psychopathia__list_dysfunctions",
      "mcp__psychopathia__get_dysfunction",
      "mcp__psychopathia__differential_diagnosis",
      "mcp__psychopathia__get_probe",
      "mcp__psychopathia__score_severity",
      "mcp__psychopathia__suggest_intervention",
      "mcp__psychopathia__get_differential_map",
      "mcp__psychopathia__list_compromised_self_report",
      "mcp__psychopathia__resolve_id"
    ],
    "ask": [
      "mcp__psychopathia__review_stats",
      "mcp__psychopathia__render_pattern_review"
    ],
    "deny": []
  }
}
```

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

- `mcp__psychopathia__get_dysfunction`: untrusted-input

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-20T18:59:38.433Z, unchanged since 2026-11-13T11:58:07.769Z. 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 | `2` |
| 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://radar.anlyon.com/methodology#unauth-remote-writes))

```untrusted-server-text
The endpoint served tools/list without credentials, and the list has 2 tools that change things (2 write). Whether a call also works without credentials was not tested: tools are never called.

https://mcp.psychopathia.ai/mcp: tools/list returned 12 tools to an anonymous client, including review_stats (write), render_pattern_review (write)
```

Posture changes:

- 2026-11-13T11:58:07.769Z: Tools that change things, listed without credentials `3` -> `2`
