Run agent evaluation on agent data. Accepts both plain text and JSON strings. Parameters: - evaluator_name: Name of the agent evaluator to use (intent_resolution, tool_call_accuracy, task_adherence) - query: User query (plain text or JSON string) - response: Agent response (plain text or JSON str...
AI agents invoke run_agent_eval to trigger actions in Azure AI Agent Service 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 executes evaluation logic on agent data based on specified evaluator types. While it does not directly modify persistent data (Write), delete data (Destructive), or move money (Financial), it performs triggered operations with external effects—running evaluations that could influence downstream decisions.
From the tool's definition Tool name 'run_agent_eval' and description 'Run agent evaluation' indicates execution of evaluation logic.
Documented attack patterns abuse exactly the kind of access run_agent_eval gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Azure AI Agent Service MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for run_agent_eval:
{
"version": "1",
"default": "deny",
"tools": {
"run_agent_eval": {
"limits": [
{
"counter": "run_agent_eval_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}run_agent_eval 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.
Run agent evaluation on agent data. Accepts both plain text and JSON strings. Parameters: - evaluator_name: Name of the agent evaluator to use (intent_resolution, tool_call_accuracy, task_adherence) - query: User query (plain text or JSON string) - response: Agent response (plain text or JSON string) - tool_calls: Optional tool calls data (JSON string) - tool_definitions: Optional tool definitions (JSON string). It is categorised as a Execute tool in the Azure AI Agent Service MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Azure AI Agent Service MCP Server MCP server in PolicyLayer and add a rule for run_agent_eval: 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 Azure AI Agent Service MCP Server. Nothing to install.
run_agent_eval 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 run_agent_eval 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 run_agent_eval. 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.
run_agent_eval is provided by the Azure AI Agent Service MCP Server MCP server (microsoft-foundry/mcp-foundry). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Azure AI Agent Service 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.
28 Azure AI Agent Service MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.