Critical Risk →

em_escrow_refund

Refund escrowed funds back to the agent (cancel task). The on-chain flow: Escrow contract -> PaymentOperator.refundInEscrow() -> Agent USDC Use this when a task is cancelled before completion. Only works if funds are still in escrow (not yet released). Args: params: task_id, optional amount (defa...

Part of the Execution Market server.

em_escrow_refund can move real money through Execution Market, with no limits today. PolicyLayer puts allow, deny, and rate-limit rules on every call. Live in minutes.

SECURE EXECUTION MARKET →

Free to start. No card required.

AI agents use em_escrow_refund to initiate financial transactions through Execution Market. Financial operations involve real money and are irreversible once processed. PolicyLayer blocks financial tools by default, requiring explicit human approval with transaction-level limits to prevent unauthorised spending.

em_escrow_refund moves real money. Without a policy, an autonomous agent could initiate transactions that drain accounts or exceed budgets. PolicyLayer blocks financial tools by default, requiring human-in-the-loop approval with configurable spending limits per transaction and per time window.

Financial tools involve real money. Block by default and require explicit human approval before enabling.

policy.json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "em_escrow_refund": {
      "deny_if": [
        {
          "conditions": [],
          "on_deny": "Requires human approval."
        }
      ]
    }
  }
}

See the full Execution Market policy for all 38 tools.

Get this rule live on your own Execution Market server in minutes. PolicyLayer enforces it on every call, before it runs.

ENFORCE ON MY EXECUTION MARKET →

View all 38 tools →

These attack patterns abuse exactly the kind of access em_escrow_refund gives an agent. Each links to the full case and the policy that stops it:

Browse the full MCP Attack Database →

Every attack above starts with a tool call. PolicyLayer checks each one against your policy first, so em_escrow_refund only ever does what you allow.

SECURE EXECUTION MARKET →

Other financial tools across the catalogue. The same approach applies to each: deny by default, or require human approval.

What does the em_escrow_refund tool do? +

Refund escrowed funds back to the agent (cancel task). The on-chain flow: Escrow contract -> PaymentOperator.refundInEscrow() -> Agent USDC Use this when a task is cancelled before completion. Only works if funds are still in escrow (not yet released). Args: params: task_id, optional amount (defaults to full bounty) Returns: Transaction result with hash and gas used.. It is categorised as a Financial tool in the Execution Market MCP Server, which means it involves financial transactions. Block by default and require explicit approval.

How do I enforce a policy on em_escrow_refund? +

Register the Execution Market MCP server in PolicyLayer and add a rule for em_escrow_refund: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Execution Market. Nothing to install.

What risk level is em_escrow_refund? +

em_escrow_refund is a Financial tool with critical risk. Critical-risk tools should be blocked by default and only enabled with explicit human approval.

Can I rate-limit em_escrow_refund? +

Yes. Add a rate_limit block to the em_escrow_refund rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.

How do I block em_escrow_refund completely? +

Set action: deny in the PolicyLayer policy for em_escrow_refund. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.

What MCP server provides em_escrow_refund? +

em_escrow_refund is provided by the Execution Market MCP server (https://mcp.execution.market/mcp/). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Execution Market tool call.

Deterministic rules across all 38 Execution Market tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

Free to start. No card required.

4,600+ MCP servers and 31,000+ tools scanned and risk-classified.

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.