← all tools
◆ AI/LLM Security
agent-safe-pipeline
v0.1.0 open-sourceReference architecture for AI agents that propose actions but cannot authorize them — immutable intent capture, an independent Decionis policy verdict (ALLOW/ESCALATE/BLOCK), verified human approval, and a SafeExecutor that consumes a single-use intent-bound grant.
Summary
Reference architecture for AI agents that propose actions but cannot authorize them — immutable intent capture, an independent Decionis policy verdict (ALLOW/ESCALATE/BLOCK), verified human approval, and a SafeExecutor that consumes a single-use intent-bound grant.
Release history
- v0.1.0
agent-safe-pipeline v0.1.0 introduces an agent-safe execution pipeline for controlled AI agent operations.
└──▷ GET THIS VERSION$ git clone --branch v0.1.0 https://github.com/decionis/agent-safe-pipeline.git # already have the repo? check out this version: $ git checkout v0.1.0
- ›Introduces an agent-safe execution pipeline for running AI agent workloads in a controlled, inspectable environment.