# `ai.dsght/public` 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/ai.dsght/public
- JSON: https://radar.anlyon.com/api/v1/servers/ai.dsght/public
- Version: 1.2.0 (active)
- Publisher: ai.dsght (domain proven to the MCP Registry)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Repository: https://github.com/michalstrnadel/dsght-landing
- Metadata from the official MCP Registry, synced 2026-10-21T17:41:40.226Z

Title and description, as the publisher wrote them:

```untrusted-server-text
Read-only MCP access to DSGHT.ai published foresight and country-grounded claim resolution.
```

Scanned 2026-11-10T02:17:49.082Z 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__public__detect_tensions` `mcp__public__extract_assumptions` `mcp__public__get_public_claims` |
| 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 (2)

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

### `detect_tensions`

- 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.
- 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
Use when the user pastes strategy text and asks what is internally inconsistent: detects the structural contradictions — where the text quietly bets on two things that cannot both hold. Stateless taster; nothing is stored. DSGHT surfaces the conflict rather than smoothing it over.
```

### `extract_assumptions`

- 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.
- 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
Use when the user pastes a strategy, plan, or memo and wants its hidden bets made explicit: extracts the falsifiable, load-bearing assumptions (dated/quantified bets + structural findings). Stateless taster — nothing is stored; the persistent per-company assumption registry with monitoring and a Brier track record is the authenticated tier.
```

## Only reads (10)

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

### `get_country_indicator`

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

Description, as the server wrote it:

```untrusted-server-text
Look up a country's authoritative macro indicator (inflation | gdp_growth) directly — deterministic, offline, cited. Use whenever you need a real macro figure instead of recalling one from memory: give the jurisdiction (CZ, DE, US, ...) and indicator, plus an optional year (omit for the full annual series).
```

### `get_public_claims`

- The name "get_public_claims" 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 evidence claims behind a published future-space that carry a cited source page (URL + quoted passage), with a census of how many claims are verified, cited or uncited. Use when you need quotable, source-backed facts on a covered topic. Uncited claims are counted, never served as sourced.
```

### `get_public_forecasts`

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

Description, as the server wrote it:

```untrusted-server-text
Get the append-only forecast registry of a published future-space — every checkable signpost with the probability, threshold and date stated in advance, its resolution and Brier score — plus the track record. Use when the user asks how reliable DSGHT's forecasts have been, or what dated, checkable predictions a space makes.
```

### `get_public_report`

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

Description, as the server wrote it:

```untrusted-server-text
Get the latest full foresight report (markdown) for a published future-space. Use when the user wants the detailed analysis, or a quotable, cited source on a covered topic.
```

### `get_public_scenarios`

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

Description, as the server wrote it:

```untrusted-server-text
Get the scenario matrix with probabilities for a published future-space. Use when the user asks how a domain (banking, energy, CEE markets, ...) might play out, what the plausible futures are, or wants to reason across scenarios instead of a single forecast.
```

### `get_public_tensions`

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

Description, as the server wrote it:

```untrusted-server-text
Get the structural tensions — the load-bearing contradictions — for a published future-space. Use when the user asks what forces conflict, what could break a strategy, or what the tensions/paradoxes are in a covered domain. DSGHT surfaces contradictions rather than averaging them.
```

### `list_countries`

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

Description, as the server wrote it:

```untrusted-server-text
List the jurisdictions and macro indicators (inflation, GDP growth) available to resolve_country_claim and get_country_indicator, with per-country year coverage and source. Call this to check coverage before resolving a claim or looking up an indicator.
```

### `list_public_spaces`

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

Description, as the server wrote it:

```untrusted-server-text
List the DSGHT future-spaces published for public use (slug, topic, status, depth). Call this first to discover what foresight is available before fetching scenarios, tensions, or a report.
```

### `resolve_country_claim`

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

Description, as the server wrote it:

```untrusted-server-text
Use this WHENEVER the user states or asks you to verify a numeric macro claim (inflation, GDP, interest rates): settle it against the claim's OWN country's authoritative series instead of answering from memory, which risks the wrong jurisdiction (e.g. U.S. CPI for a Czech claim). Binds the claim to a jurisdiction (CZ, DE, ...) and returns a cited above/below verdict. Deterministic and offline.
```

### `search_spaces`

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

Description, as the server wrote it:

```untrusted-server-text
Find the most relevant published DSGHT future-space by keyword (topic + report text, plain-text ranked, no AI). Use this to locate a space before calling get_public_scenarios / get_public_tensions / get_public_report when you do not already know the slug.
```

## Claude Code permissions

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

```json
{
  "permissions": {
    "allow": [
      "mcp__public__list_public_spaces",
      "mcp__public__get_public_tensions",
      "mcp__public__get_public_scenarios",
      "mcp__public__get_public_report",
      "mcp__public__get_public_claims",
      "mcp__public__get_public_forecasts",
      "mcp__public__search_spaces",
      "mcp__public__list_countries",
      "mcp__public__get_country_indicator",
      "mcp__public__resolve_country_claim"
    ],
    "ask": [
      "mcp__public__extract_assumptions",
      "mcp__public__detect_tensions"
    ],
    "deny": []
  }
}
```

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

- `mcp__public__get_public_claims`: untrusted-input
- `mcp__public__extract_assumptions`: untrusted-input
- `mcp__public__detect_tensions`: untrusted-input

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-20T18:25:45.466Z, unchanged since 2026-10-22T14:19:44.832Z. 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 | `2025-11-25` |
| 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://www.dsght.ai/mcp: tools/list returned 12 tools to an anonymous client, including extract_assumptions (write), detect_tensions (write)
```
