# `wiki.agenticweb/probe` 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/wiki.agenticweb/probe
- JSON: https://radar.anlyon.com/api/v1/servers/wiki.agenticweb/probe
- Version: 1.3.0 (active)
- Publisher: wiki.agenticweb (domain proven to the MCP Registry)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Repository: https://github.com/Belgravia-Labs/agenticweb-data
- Metadata from the official MCP Registry, synced 2026-10-25T18:27:38.651Z

Title and description, as the publisher wrote them:

```untrusted-server-text
Agentic Web Wiki — standards reference, probe, validate, template
Look up 133 agentic-web standards, live-probe any domain, validate manifests, generate boilerplate.
```

Scanned 2026-11-09T06:18:06.491Z over streamable-http.

**11 tools: 5 change things and 6 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 0 | deny |
| Can't be undone | irreversible | 0 | ask |
| Changes things | write | 5 | ask |
| Only reads | read | 6 | 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__probe__digest` `mcp__probe__has` `mcp__probe__probe` `mcp__probe__validate_manifest` |
| 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 `template` ([annotation-conflict](https://radar.anlyon.com/methodology#annotation-conflict))

```untrusted-server-text
The server marks "template" 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 "generate": a word for a change.
```


## Changes things (5)

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

### `ask`

- 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
Recommend and rank standards to implement for an engineering goal (pass question/q and/or use_case + audience filters), or compare named protocols head-to-head when the query names 2+ standards. Returns { use_case, audience, standards (ranked by tier and adoption), starter_bundle, comparisons }. Use ask for architecture and stack decisions; use search for keyword lookup or get for one standard's record.
```

### `define`

- 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
Define an agentic-web glossary term or standard by id, name, or alias. Returns { kind, id, name, definition_md, related_protocols, url } (or nearest-term hints on unknown-term error). Use define for concept definitions; use get when you need a standard's full specification record, or search for keyword discovery.
```

### `has`

- 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
Check one public origin over HTTPS for a single standard or manifest (rate-limited to 120/min). Returns { origin, id, found, outcome, status, url, reason, article } without grading the full site. Use has for a fast single-file check; use probe to scan all 37 paths at once, or validate_manifest for A+–F conformance grading.
```

### `probe`

- 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
Scan a public origin across 37 agent-facing paths and headers (24h KV cache by default; pass fresh: true to re-probe live with a 5-min per-origin cooldown, rate-limited to 30/min). Returns { found_ids, found, absent, coverage, extracted, connect_recipe }; pass brief: true for found_ids + extracted endpoints only, compact: true to replace absent/skipped maps with absent_ids[], or verify: true to handshake MCP/A2A endpoints. Use probe for raw multi-path discovery; use has for one standard, or validate_manifest for A+–F conformance scoring.
```

### `template`

- The description's first verb is "generate": 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
Generate in-memory starter boilerplate for any of the 12 supported manifest ids (ard, ai-catalog-well-known-file, llms-txt, agents-md, anthropic-agent-skills-format, mcp-server-cards, a2a, ai-crawler-tokens-robots, api-catalog-rfc9727, security-txt, auth-md, okf) without writing to disk. Returns { id, path, content_type, contents, validate_with }; when from_live: true is passed with domain, it probes the public origin over HTTPS to pre-fill discovered MCP, OpenAPI, A2A, and contact URLs. Use template to author a new manifest; use validate_manifest to lint existing content.
```

## Only reads (6)

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

### `changes`

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

Description, as the server wrote it:

```untrusted-server-text
List machine-verified specification changes since a YYYY-MM-DD date or ISO-8601 timestamp (newest first, default 50, max 200). Returns { since, count, entries: [{ date, record_id, field, old, new, source }] } filtered optionally by lifecycle step or record_id. Use changes for field-level spec diffs; use stale to audit unverified records or digest for daily ecosystem news.
```

### `digest`

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

Description, as the server wrote it:

```untrusted-server-text
Fetch the curated daily agentic-web news and adoption digest for a YYYY-MM-DD date (defaults to latest). Returns { date, headline, highlights, new_instances, spec_changes, stats }; pass list: true to return available dates only, or compact: true (alias: include_stats: false) to omit the census stats block. Use digest for daily narrative summaries; use changes for raw field-level diffs.
```

### `get`

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

Description, as the server wrote it:

```untrusted-server-text
Fetch one standard's complete specification record by id, acronym, filename, or RFC number. Returns { id, name, step, status, governance, version, spec_url, probe, adoption, summary_md, url } (or unknown-id error with nearest matches). Use get for a known standard; use search to find unknown ids or define for glossary concepts.
```

### `search`

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

Description, as the server wrote it:

```untrusted-server-text
Search 132 standards and 52 glossary terms by keyword, acronym, alias, or 'A vs B' pair (pass q or alias query, 2–200 chars). Returns { results (top 10 ranked by title/alias/body match), matching_terms, comparison_url } or nearest_ids hints when empty. Use search for keyword discovery; use get for an exact id or ask for stack recommendations.
```

### `stale`

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

Description, as the server wrote it:

```untrusted-server-text
List standards whose last_verified date is older than N days (default 30, oldest first), optionally filtered by lifecycle step. Returns { threshold_days, count, records: [{ id, name, step, last_verified, days_since_verified }] }. Use stale to audit verification freshness; use changes to inspect actual specification edits since a date.
```

### `validate_manifest`

- The name "validate_manifest" says "validate": a word for reading.
- 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
Audit a public origin across 37 agent manifests (24h KV cache; fresh: true re-probes live with 5-min cooldown, 30/min limit) or validate local file text in-memory before deploy (pass id + content, no network fetch). Returns { grade (A+ to F), score, subscores, conformance_summary, connect_recipe, permissions, manifests }; combine verify: true (live MCP/A2A handshakes), bridge: true (ARD cross-manifest check), diff: true (score_delta vs prior snapshot), or brief/compact payload trimming. Use validate_manifest for conformance grading and local linting; use probe for ungraded discovery or template to generate starter files.
```

## Claude Code permissions

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

```json
{
  "permissions": {
    "allow": [
      "mcp__probe__get",
      "mcp__probe__search",
      "mcp__probe__changes",
      "mcp__probe__stale",
      "mcp__probe__validate_manifest",
      "mcp__probe__digest"
    ],
    "ask": [
      "mcp__probe__define",
      "mcp__probe__has",
      "mcp__probe__probe",
      "mcp__probe__template",
      "mcp__probe__ask"
    ],
    "deny": []
  }
}
```

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

- `mcp__probe__has`: untrusted-input
- `mcp__probe__probe`: untrusted-input
- `mcp__probe__validate_manifest`: untrusted-input
- `mcp__probe__digest`: untrusted-input

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-20T19:21:01.700Z, unchanged since 2026-11-20T19:21:01.700Z. 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 | `5` |
| 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 5 tools that change things (5 write). Whether a call also works without credentials was not tested: tools are never called.

https://agenticweb.wiki/mcp: tools/list returned 13 tools to an anonymous client, including define (write), has (write), probe (write), template (write), ask (write)
```

Posture changes:

- 2026-11-20T19:21:01.700Z: RFC 9728 protected resource metadata `https://agenticweb.wiki/.well-known/oauth-protected-resource/mcp` -> `none`; Authorization servers `https://agenticweb.wiki` -> `none`; Dynamic client registration `yes` -> `not observed`; Client ID metadata documents `no` -> `not observed`
- 2026-10-29T22:17:54.964Z: Tools that change things, listed without credentials `8` -> `5`; RFC 9728 protected resource metadata `none` -> `https://agenticweb.wiki/.well-known/oauth-protected-resource/mcp`; Authorization servers `none` -> `https://agenticweb.wiki`; Dynamic client registration `not observed` -> `yes`; Client ID metadata documents `not observed` -> `no`
