Checks / Agent output: number sanity / Score deltas are monotonic
Agent output: number sanity
Score deltas are monotonic
This control asks you to show: Score deltas are monotonic. Checks that numbers an AI produced are within sensible bounds and internally consistent.
What evidence passes
The check reads the field
score_history in what the AI step produced. It passes when that field is a value that only ever goes up. 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_math_gate_055", "output_payload": {"score_history": "..."}}Or run every Agent output: number sanity control at once with sebbi_run_framework. How to connect