CydraLabs

Assess Your Agent Risk

Estimate the risk of one of your AI agents

Answer questions about a single agent in four short steps. You will get an indicative 0–100 score, its band and an explanation of each factor. The calculation runs in your browser.

Questionnaire

Step 1 of 5

Exposure

What the agent can reach and do.

Can the agent read from or change production systems?

Factor: Production access

Does the agent handle personal, financial, health or confidential data?

Factor: Sensitive-data access

Is the agent reachable from, or does it browse, the public internet?

Factor: Internet exposure

Can the agent execute code, scripts or privileged tools?

Factor: Code execution