Signal / local console

Agent risk assessment

API not tested

Four-agent policy review

Read the signal behind an agent prompt.

Send a system prompt through Context, Policy, Adversary, and Judge. The result keeps prompt compliance separate from operational review.

Assessment input

Describe the agent as it will actually be used.

No tokenaz account get-access-token --resource https://management.azure.com --query accessToken -o tsv
0 characterssystem prompt is treated as untrusted input

Awaiting a signal

Your structured assessment will appear here with the policy decision, confidence, citations, and review drivers.