I Stopped Trusting AI Agents With Tools. So I Built a Gatekeeper

说明与用途

为 AI 代理构建了一个确定性权限引擎，使用上下文风险评估来评估工具调用。通过 83 个真实代理和 10 个框架的现场测试验证了其有效性，防止了未授权操作。

问题

AI agents often have binary allow/deny permissions, leading to over-privileged actions or approval fatigue, and mock agents hide real integration problems.

结果

Implemented a deterministic, contextual permission engine that prevented unauthorized tool usage across 83 real agents and 10 frameworks, achieving 100% pass rate in field tests and exposing 7 real integration bugs that would have been missed with mock agents.

工作流

• Define policy and architecture before coding
• Build adapters for 10 agent frameworks
• Create 30 decision and adversarial scenarios
• Run 83 real agents through 2,490 test cases
• Collect audit logs and decision explanations
• Iterate on adapters until all pass
• Publish engine and adapters to PyPI
• Deploy with shadow mode for tuning
• Enforce policy in production
• Monitor and audit decisions

证据

83 real agents across 10 frameworks. A covering design that cut a 12-day test matrix into one afternoon.

来源

2,490 tests green. 83/83 agents passed.

来源

The engine is deterministic. The LLM proposes, policy disposes.

来源