Judgment infrastructure
for consequential actions.
We are building Nimclea Orange, a deterministic authority gateway for AI agent workflows.
The Orange Executive Brief is available.
Built from real-world judgment problems.
Nimclea Orange grew from our early work on plastic-waste decisions, where evidence quality, authority boundaries, and release conditions could not be treated as afterthoughts.
Orange is not merely an evidence pack or management dashboard. It sits before consequential actions, evaluates authority and qualified evidence, and produces either an Issue Card or a Release Log.
Observable Events
Real-world actions begin with field conditions, records, and measurable facts, not unverified claims.
Qualified Evidence
Inputs must be structured, bounded, and reviewable before they can support a consequential decision.
Authority Before Release
A decision is not released until authority, evidence, and boundary conditions are satisfied.
Orange is Trust Customs for AI agents.
A prompt is not authority.
Agents may write code.
Claims are not evidence.
Agents may call tools.
Capability is not permission.
Agents may change reality.
Agents can act. But agents cannot issue their own clearance.