Building open infrastructure for reliable and governable agentic systems.
Enkronos develops open-source tooling and applied software for modular AI systems, with a focus on reliability, delegation, governance, and production readiness.
Our current open-source modules include:
• Agent Contracts — task delegation contracts • CapToken — capability authorization tokens • GuardMesh — runtime policy enforcement • TraceForge — portable execution traces • SkillPulse — agent skill reliability • AgentEvalOps — failure-mode evaluation
We are particularly interested in AI systems that are:
- reliable under change
- explicit in their interfaces
- modular in their capabilities
- easier to inspect and govern
- useful in real operational environments
Our current work explores topics such as:
- skill health and degradation detection
- contract-based delegation
- agent governance layers
- traceable execution patterns
- portable infrastructure for agentic systems
Check the pinned repositories below for the latest public projects.