liveBook · Manning

Cerbos

As explained in AI Governance

Cerbos is a policy engine for evaluating authorization requests against declarative access-control policies.

Cerbos is a deterministic authorization policy engine for agent tool access. It evaluates each tool call against human-readable policies and returns allow or deny, independently of the model’s reasoning. section 4.9

Authorization in a claims assistant

An agent investigating a denied claim can query claim records, adjuster notes, and the appeals process, then draft an explanation. Cerbos governs whether its tool calls are allowed. section 4.9

How AI Governance covers it

How deep it goes
Defines deterministic per-call allow-or-deny checks against human-readable policies without detailing policy syntax or engine configuration.
Examples
Uses a claims agent that queries records, adjuster notes, and appeals information before drafting a denial explanation.
What you can do after
Apply Cerbos to allow or deny an agent’s tool calls independently of its reasoning.
How it connects
Connects authorization policies to agent tool calls while separating access decisions from model reasoning.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Governance. The section links point to the parts of the book it draws on.