agent-safety-frameworks
Formal safety framework for AI agents with provable guarantees
Safety guardrails for LLM agents — budget enforcement, invariants, and provable guarantees. Zero dependencies.
