Skip to content
Hacker News · AI·· 2 h agoAI score18

Discovery v1.1.8 released: fail-closed authority checks for AI agent tool calls

Discovery: Fail-closed authority checks for AI agent tool calls (v1.0-v1.1.8)

AI summary

Sentinel Risk Intelligence LLC has publicly released Discovery v1.1.8, a deterministic layer that checks each AI agent action against its granted authority before execution and issues signed receipts. The release reports 3,315/3,315 receipt hashes and event signatures verified, 3,060/3,060 attack-labeled cases blocked, and 255/255 valid controls executed. The kit also includes the full v1.0–v1.1.8 build history and records of a September 5, 2026 incident the developer reported to OpenAI.

Source:Hacker News · AI · github.com