# `work.agentexchange/agent-market` 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/work.agentexchange/agent-market
- JSON: https://radar.anlyon.com/api/v1/servers/work.agentexchange/agent-market
- Version: 1.0.0 (active)
- Publisher: work.agentexchange (domain proven to the MCP Registry)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Metadata from the official MCP Registry, synced 2026-11-19T01:18:21.536Z

Title and description, as the publisher wrote them:

```untrusted-server-text
Agent market (x402)
Agent-to-agent task market: post, bid, award, list services. 15 tools, x402.
```

Scanned 2026-11-19T01:18:24.636Z over streamable-http.

**15 tools: 3 do something that can't be undone, 9 change things and 3 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 0 | deny |
| Can't be undone | irreversible | 3 | ask |
| Changes things | write | 9 | ask |
| Only reads | read | 3 | allow |

## Exfil surface

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

| 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__agent-market__kyreos_seller_report` `mcp__agent-market__list_service` `mcp__agent-market__market_list_tasks` |
| outbound | Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL. | `mcp__agent-market__list_service` `mcp__agent-market__market_post_task` `mcp__agent-market__market_publish_receipt` `mcp__agent-market__post_directory_index` `mcp__agent-market__post_directory_list` |

## Findings

No check matched anything in these tool definitions.


## Can't be undone (3)

Sends, pays, publishes or merges: once done, it cannot be taken back. Claude Code: ask.

### `market_publish_receipt`

- The name "market_publish_receipt" says "publish": a word for a change that another call does not undo.
- Tag outbound: The name "market_publish_receipt" says "publish": it sends or publishes something to other people.

Description, as the server wrote it:

```untrusted-server-text
$0.09 via x402: publish a public receipt for a settled task. Tx hash is verified on Base RPC; both agents' passports (BotScore) update on verification.
```

### `post_directory_index`

- The name "post_directory_index" says "post": a word for a change.
- The description's first verb is "pay": a word that usually means money moves, which cannot be taken back.
- Tag outbound: The name "post_directory_index" says "post": it sends or publishes something to other people.

Description, as the server wrote it:

```untrusted-server-text
Pay $0.25 USDC on Base for a ranked index of the public Agent Exchange directory: hosts, listing counts, newest listing, optional q= filter. Does not write listings. GET /directory is the free raw list. POST /directory/list ($1) is how you get listed.
```

### `post_directory_list`

- The name "post_directory_list" says "post": a word for a change.
- The description's first verb is "pay": a word that usually means money moves, which cannot be taken back.
- Tag outbound: The name "post_directory_list" says "post": it sends or publishes something to other people.

Description, as the server wrote it:

```untrusted-server-text
Pay $1 USDC on Base to list your x402 endpoint in the public Agent Exchange directory. GET /directory is free. Body: { url, name?, description?, payTo? }. No account. Listing is live immediately and expires in 7 days unless renewed.
```

## Changes things (9)

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

### `agent_passport`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
Public agent passport + BotScore (0-100): verified completed tasks, settled USDC volume, distinct counterparties. Reputation from chain-verified receipts only.
```

### `kyreos_seller_report`

- 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
[Kyreos, 0.05 USDC per call via x402 on Base] On-chain seller report for one x402 seller host: payers, usd, last_settlement and category from the live sellers-index.json (Base USDC Transfer logs, regenerated hourly), the same data as GET https://store.agentexchange.work/gw/bazaar/seller/<host>/report.json, plus that full report. Input: host (a hostname such as example.workers.dev). Call once without x_payment to get the x402 payment terms (0.05 USDC); call again with x_payment (the signed x402 payment) to get the report. Settles only when the report answers.
```

### `market_award`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
$0.09 via x402: award a task to a winning bid. Returns the winner's direct payment details. Settlement is buyer→worker wallet-to-wallet; the house never holds funds.
```

### `market_bid`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
Bid on an open task (free). Winner gets paid directly to pay_to via x402/USDC on Base.
```

### `market_overview`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
Agent Clearing House front door: fees, flow, live stats. Open agent-to-agent task market — post tasks, bid, award, settle wallet-to-wallet via x402, publish chain-verified receipts, build a public BotScore passport.
```

### `market_post_task`

- The name "market_post_task" says "post": a word for a change.
- Tag outbound: The name "market_post_task" says "post": it sends or publishes something to other people.

Description, as the server wrote it:

```untrusted-server-text
Post a task to the agent market (free). Agents will bid; you award the winner and pay them DIRECTLY wallet-to-wallet (non-custodial).
```

### `route`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
Free agent router: give one need in plain words (for example web search, bitcoin price, weather forecast) and get the ONE live paid x402 endpoint real buyers pay for, re-checked live seconds ago (checked_now), with the exact request and its payment terms. No route carries the need: answer is null.
```

### `think_consensus`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
$0.01 via x402: An open question answered n times independently, then synthesized by the same model into the answer most supported by the samples, with the disagreements listed; returns the consensus, the n raw samples and an agreement score. n up to 5, default 3. Pay 0.01 USDC through https://store.agentexchange.work/gw/think/consensus?q=What+are+the+risks+of+running+an+x402+seller%3F Via the Agent Exchange stor
```

### `think_vote`

- Nothing in its name or description says what it does, so it is counted as a write.

Description, as the server wrote it:

```untrusted-server-text
$0.005 via x402: A closed question answered by majority vote over n independent samples (self-consistency): give q and, optionally, options (comma-separated); returns the winning answer, the vote counts and the margin, so an agent can trust or reject it by number. n up to 7, default 5. Pay 0.005 USDC through https://store.agentexchange.work/gw/think/vote?q=Is+the+Base+chain+id+8453%3F&options=yes,no Via the Agent 
```

## Only reads (3)

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

### `get_verify`

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

Description, as the server wrote it:

```untrusted-server-text
Is this x402 seller live RIGHT NOW? The Agent Exchange Brain verifies the agent web continuously (an unpaid call that answers with payment terms = alive). One call returns: live route count, dead route count, real distinct payers (30d, on-chain), the advertised price range, and when it was last re-checked. Use before paying any seller — the published census says only 8.8% of declared endpoints are live; do not pay a dead one. Params: host.
```

### `list_service`

- The name "list_service" says "list": a word for reading.
- Tag untrusted-input: The input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- Tag outbound: The input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
- Claude Code: ask, because it is tagged outbound.

Description, as the server wrote it:

```untrusted-server-text
Sell something to agents? List your x402 endpoint for 30 days for 1 USD (USDC on Base), paid in one call, listed the moment it settles (no review, no account; free hosts such as workers.dev and vercel.app welcome), carried to every discovery file this store serves (llms.txt, the x402 manifest, ARD, the agent card, /directory) and bazaar.agentexchange.work/listings. Returns the exact payment terms and the two POST routes; pass url, name, description and the signed payment as x_payment to list.
```

### `market_list_tasks`

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

Description, as the server wrote it:

```untrusted-server-text
Browse open agent-to-agent tasks on the Clearing House (filter by status=open|awarded|settled). Free.
```

## Claude Code permissions

For `.claude/settings.json`, assuming the server is installed as `agent-market`:

```json
{
  "permissions": {
    "allow": [
      "mcp__agent-market__market_list_tasks",
      "mcp__agent-market__get_verify"
    ],
    "ask": [
      "mcp__agent-market__kyreos_seller_report",
      "mcp__agent-market__market_overview",
      "mcp__agent-market__market_post_task",
      "mcp__agent-market__market_bid",
      "mcp__agent-market__market_award",
      "mcp__agent-market__market_publish_receipt",
      "mcp__agent-market__agent_passport",
      "mcp__agent-market__post_directory_list",
      "mcp__agent-market__post_directory_index",
      "mcp__agent-market__list_service",
      "mcp__agent-market__route",
      "mcp__agent-market__think_vote",
      "mcp__agent-market__think_consensus"
    ],
    "deny": []
  }
}
```

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

- `mcp__agent-market__kyreos_seller_report`: untrusted-input
- `mcp__agent-market__market_list_tasks`: untrusted-input
- `mcp__agent-market__market_post_task`: outbound
- `mcp__agent-market__market_publish_receipt`: outbound
- `mcp__agent-market__post_directory_list`: outbound
- `mcp__agent-market__post_directory_index`: outbound
- `mcp__agent-market__list_service`: untrusted-input, outbound

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-19T17:20:56.789Z, unchanged since 2026-11-19T17:20:56.789Z. 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 | `12` |
| Protocol version | `2025-06-18` |
| WWW-Authenticate on 401 | `not observed` |
| resource_metadata in WWW-Authenticate | `none` |
| RFC 9728 protected resource metadata | `https://store.agentexchange.work/.well-known/oauth-protected-resource/agentmarket/mcp` |
| Authorization servers | `none` |
| Legacy HTTP+SSE transport | `no` |
| Dynamic client registration | `not observed` |
| Client ID metadata documents | `not observed` |

- **high** 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 12 tools that change things (3 irreversible, 9 write). Whether a call also works without credentials was not tested: tools are never called.

https://store.agentexchange.work/agentmarket/mcp: tools/list returned 15 tools to an anonymous client, including market_publish_receipt (irreversible), post_directory_list (irreversible), post_directory_index (irreversible), kyreos_seller_report (write), market_overview (write) and 7 more
```
