Checks / Agent output: compliance gates / accuracy metric recorded
Agent output: compliance gates · Art.15
EU AI Act Art.15 - accuracy metric recorded
This control asks you to show: accuracy metric recorded. Checks run on what an AI agent produced, before it is acted on, for the evidence a regulator expects.
What evidence passes
The check reads the field
accuracy in what the AI step produced. It passes when that field is a number between 0 and 1. Anything else fails, with the reason stated.Check it now
Free, no account. You get PASS or FAIL, the reason, and a tamper-evident receipt.
Run it from your AI assistant
Connect
https://sebbi.pro/mcp to Claude, ChatGPT, Cursor or any MCP app, then call:sebbi_run_check {"check": "sebbi_compliance_gate_007", "output_payload": {"accuracy": "..."}}Or run every Agent output: compliance gates control at once with sebbi_run_framework. How to connect