agent-build-gates is an MIT-licensed Python package for validating code produced by AI agents. It includes anti-simulation and AWS-account tripwires plus an evaluation harness with confidence intervals and significance testing for developers and AI engineering teams.
In the Agent evaluation & testing space, agent-build-gates takes a focused approach. It focuses on validating that AI-agent-built code is genuine, accesses required AWS resources, and meets evaluation standards. agent-build-gates is an open-source project aimed at developers building and evaluating AI agents. The project is open source (MIT). agent-build-gates is available on the command line, and it can be self-hosted.
Paul Snyman builds and maintains agent-build-gates, and it first shipped in 2026. Development happens publicly on GitHub with 53 commits in the last 90 days. Key capabilities include anti-simulation tripwire, AWS-account tripwire, and eval harness. It exposes integrations via a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do