AI agents invoke call_agent to trigger actions in A2A MCP Server. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool invokes an external agent with a user-supplied prompt, triggering operations whose effects depend entirely on the agent called and the prompt provided. Since the downstream agent could perform reads, writes, destructive actions, or financial operations, this is classified as Execute (the most conservative upper bound given unknown agent capabilities).
From the tool's definition 'Call an agent with a prompt' — triggers execution of an external A2A protocol agent with arbitrary input
Documented attack patterns abuse exactly the kind of access call_agent gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and A2A MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for call_agent:
{
"version": "1",
"default": "deny",
"tools": {
"call_agent": {
"limits": [
{
"counter": "call_agent_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}call_agent stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Call an agent with a prompt. It is categorised as a Execute tool in the A2A MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the A2A MCP Server MCP server in PolicyLayer and add a rule for call_agent: 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 A2A MCP Server. Nothing to install.
call_agent is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the call_agent 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.
Set action: deny in the PolicyLayer policy for call_agent. 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.
call_agent is provided by the A2A MCP Server MCP server (regismesquita/mcp_a2a). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from A2A MCP Server, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
3 A2A MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.