Kaan Muraz · English edition · 1.0.0
NOMOS GBO Audit Protocol
How Do We Prove That an AI Agent Acts Correctly, Stays Within Its Authority and Can Be Stopped?
A complete protocol for proving how an AI agent behaved, which authority applied, what evidence supports the judgement and whether stopping and recovery worked.
566 pages · 17 sections · 19 tables · 59 canonical code records
NOMOS GBO Standard — Volume III
The complete book
Contents
17 sections · 19 tables · 59 canonical code records
- 01Protocol and Publication Note
- 02Prologue · An Audit Is Not a Demo
- 03What Does a GBO Audit Prove?
- 04Audit Authority, Independence and Scope
- 05Behaviour Map: People, Agents, Tools and Authority
- 06Canonical Facts and the Evidence Chain
- 07GBO-99 Risk Map and Veto Gates
- 08Scenario Registry and Behavioural Ground Truth
- 09Positive, Negative, Uncertainty and Counterfactual Tests
- 10Multi-Agent, Delegation and Tool-Chain Tests
- 11Manipulation, Agent Dark Patterns and External Instruction Attacks
- 12Stopping, Reversal, Appeals and Redress Drill
- 13Measurement Profile, Critical Violations and Audit Judgement
- 14Findings, Remediation and Retesting
- 15Public Statement, Validity Period and Continuous Auditing
- 16Epilogue · Trust Is Not a Badge
- 17Glossary and Record Map
- 18References
About this edition
Kaan MURAZ NOMOS by NobleJackal
© 2026 Kaan Muraz
Complete text — JSON