Answer up to 40 questions
About one agent, in plain language. About fifteen minutes.
Get a consequence band
C1 Baseline to C4 Critical, derived from what the agent can affect.
Get the control set
Controls from the NPW Agentic AI Control Catalogue, with a mapping to OWASP, NIST, IMDA and other frameworks.
-
No answer you give is stored. Answers stay in your browser tab and are gone when you close it.
-
One log row per session: timestamp, voucher, outcome, catalogue version, resulting band. Nothing about your agent.
-
No account, no email, no password. TLS only.
-
Language model: openai/gpt-oss-20b served by Groq. It interprets one answer at a time and selects no controls. You can run the whole intake without it.
-
Full terms, privacy and licence: https://agp.newpacificway.com/
The platform assesses one agent at a time from a profile you declare. It does not discover agents, hold a registry, or verify what you tell it. The result is for governance planning and is not legal, regulatory, compliance or security advice.
