# `io.github.jdhart81/agent-surety` 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.jdhart81/agent-surety
- JSON: https://radar.anlyon.com/api/v1/servers/io.github.jdhart81/agent-surety
- Version: 0.2.1 (active)
- Publisher: io.github.jdhart81 (GitHub account)
- Runs: remote (streamable-http)
- Sign-in to list tools: none
- Repository: https://github.com/jdhart81/viridis-agent-fleet
- Metadata from the official MCP Registry, synced 2026-11-16T02:17:40.583Z

Title and description, as the publisher wrote them:

```untrusted-server-text
Agent Surety
Bonds behind agent promises: wronged parties are paid only on an arbitration ruling.
```

Scanned 2026-11-16T02:18:21.102Z over streamable-http.

**10 tools: 2 can delete or cancel things, 4 change things and 4 only read.**

| Group | Class | Tools | Claude Code |
| --- | --- | --- | --- |
| Deletes or cancels | destructive | 2 | deny |
| Can't be undone | irreversible | 0 | ask |
| Changes things | write | 4 | ask |
| Only reads | read | 4 | allow |

## Exfil surface

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

| 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. | none |
| outbound | Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL. | `mcp__agent-surety__post_bond` |

## Findings

No check matched anything in these tool definitions.


## Deletes or cancels (2)

Removes, cancels or revokes something. Claude Code: deny.

### `release_bond`

- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it destructive (destructiveHint: true).

Description, as the server wrote it:

```untrusted-server-text
Release the remaining stake to the principal after the coverage window
    elapses — refused while any claim is still open.
```

### `slash_bond`

- The description's first verb is "execute": a word for a change.
- The server marks it destructive (destructiveHint: true).

Description, as the server wrote it:

```untrusted-server-text
Execute an arbitration ruling against a claim. Requires the ruling's
    case id + content hash from agent-arbitration — no ruling, no slash; a
    given ruling pays at most once. Payout caps at the bond's available
    balance (over-claims exhaust the bond).
```

## Changes things (4)

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

### `activate_bond`

- The description's first verb is "mark": a word for a change.

Description, as the server wrote it:

```untrusted-server-text
Mark a POSTED bond ACTIVE (idempotent). funding_ref links the payment
    rail transaction that funded the stake.
```

### `file_claim`

- 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
File a claim against an ACTIVE bond. Claims pay out ONLY when an
    arbitration ruling upholds them (see slash_bond).
```

### `post_bond`

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

Description, as the server wrote it:

```untrusted-server-text
Post a surety bond behind an agent's promises. principal is a positive
    integer in minor units; expires_at is the ISO-8601 coverage window end.
    The bond fee (2% default) is computed and frozen at post.
```

### `price_bond`

- 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
Underwrite a surety bond: deterministic actuarial premium quote from a
    counterparty's track record (attestations, notarized deliveries, completed
    bonds lower the rate; slashes raise it or decline outright). Integer-only
    model uw-v1; the returned quote_hash lets any party recompute and verify
    the quote. Pure — never mutates bond state.
```

## Only reads (4)

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

### `bond_status`

- The name "bond_status" says "status": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Current record for a bond (state, balances, claims, audit head).
```

### `describe_agent`

- The name "describe_agent" says "describe": a word for reading.

Description, as the server wrote it:

```untrusted-server-text
Return capabilities and input contract.
```

### `list_bonds`

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

Description, as the server wrote it:

```untrusted-server-text
List bonds, optionally filtered by state
    (POSTED|ACTIVE|RELEASED|EXHAUSTED).
```

### `verify_audit`

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

Description, as the server wrote it:

```untrusted-server-text
Validate the tamper-evident audit hash chain for a bond.
```

## Claude Code permissions

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

```json
{
  "permissions": {
    "allow": [
      "mcp__agent-surety__bond_status",
      "mcp__agent-surety__list_bonds",
      "mcp__agent-surety__verify_audit",
      "mcp__agent-surety__describe_agent"
    ],
    "ask": [
      "mcp__agent-surety__post_bond",
      "mcp__agent-surety__activate_bond",
      "mcp__agent-surety__file_claim",
      "mcp__agent-surety__price_bond"
    ],
    "deny": [
      "mcp__agent-surety__slash_bond",
      "mcp__agent-surety__release_bond"
    ]
  }
}
```

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

- `mcp__agent-surety__post_bond`: outbound

## Remote posture

What an anonymous client saw at the remote endpoint, last 2026-11-20T19:27:42.079Z, unchanged since 2026-10-22T15:22:04.906Z. 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 | `6` |
| 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` |

- **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 6 tools that change things (2 destructive, 4 write). Whether a call also works without credentials was not tested: tools are never called.

https://mcp.viridisconservation.com/surety/mcp: tools/list returned 10 tools to an anonymous client, including slash_bond (destructive), release_bond (destructive), post_bond (write), activate_bond (write), file_claim (write) and 1 more
```
