Connecting an agent usually starts a review: which account it uses, what it can see, whether it can change anything, and where the data goes. Every answer is another meeting.
In Settings, pick your client and Pathix writes the config block; there's no client ID to copy, because the client registers itself the first time it connects. Then you sign in, as you, with your Microsoft Entra account: no service account, no shared key. The agent sees what you see in Pathix, and nothing more.
Now it can just ask what writes to account credit limit. One tool call returns the same thirteen writers as Ep. 01, with the tier on every edge, so your agent can tell what was proved from what was proposed.
Ask it to dismiss an AI-derived edge, and it can't: no Pathix tool changes your environment, or the triage inside Pathix, and validating stays a person's decision, in the AI Audit hub. It never even calls Dataverse, and Pathix never stored a record value, so it can't answer one account's credit limit either. For data, your agent holds the Dataverse MCP; most agents that need both hold both.
A read-only metadata scan of your environment that reads the logic inside the components: compiled plugins, flow definitions, workflow XAML and scripts.
How it works →A 30-minute walkthrough on a pre-scanned demo environment. No access to your tenant, nothing to install.