Getting started

Step 1

Sign in to your workspace

AgentProof signs you in with a one-time magic link emailed to you (sent from [email protected]) - there is no password by default. If the link does not arrive, check your spam, junk, and quarantine folders and request a fresh one. See the Sign-in and magic links article for expired-link, wrong-email, and sign-in-loop fixes.

Step 2

Open the environment setup wizard

Choose how your agents live — Microsoft Power Platform / Copilot Studio, Azure AI Foundry, OpenAI, Anthropic, Google Vertex, AWS Bedrock, ServiceNow, Salesforce, open-source framework, or manual fallback. The wizard handles connector selection without needing technical setup notes.

Step 3

Add an agent under the environment

Discovered agents appear automatically when a connector is wired. Manual fallback is available for environments that have no connector path - to add one yourself, or if an environment shows none, see the No agents found help at /help/no-agents-found.

Step 4

Run an assessment

Open the environment assessment dashboard. Confirm the few unresolved facts for each agent - anything already known is prefilled and stays editable - then complete only the follow-up packs that apply. Evidence text is captured alongside each answer.

Step 5

Generate a report and start improvements

The report ends with improvement actions. Mark actions as in progress, completed, accepted risk, or deferred (new actions start as open). Reassessment is recommended when Radar intelligence updates the picture.