The UAI-1 capability ladder is an evidence-routing ladder, not an authorization ladder. Levels L0 through L6 describe what a receiver can reliably read, validate, propose, coordinate, or review before local policy decides whether any action happens.
| Level | Label | UAIX expectation | No-op behavior |
|---|---|---|---|
| L0 | Rule-Based Workflow / static format follower | Reads static public text, Markdown, sitemap entries, and small JSON. No JavaScript, POST, tool, credential, memory, or execution authority is implied. | No-op plus human review when schema, authority, or route intent is ambiguous. |
| L1 | Basic Responder / prompt-bound assistant | Fetches public JSON and follows explicit prompt-bound instructions. It cannot approve writes or private actions. | Stop with human_review_required when requested work needs authority not present in the packet. |
| L2 | Tool User / approval-bound co-pilot | Understands schemas and can prepare tool-call proposals, but write-capable behavior requires explicit human approval. | Prepare no-op justification when approval, integrity, or provenance is missing. |
| L3 | Bounded Operator / autonomous within explicit limits | Operates inside stated limits when authority, policy, timeout, provenance, and evidence are explicit. | Stop on boundary conflict, missing checksum, credential validation request, or private-network probe request. |
| L4 | Domain Actor / full-domain autonomous worker under boundaries | Handles a bounded domain across a workflow with trace, review checkpoints, typed errors, and final evidence. | Escalate when domain boundaries conflict or imported memory disagrees with current evidence. |
| L5 | Multi-Agent Collaborator / peer coordination with evidence handoff | Coordinates multiple agents while UAIX carries portable evidence, role boundaries, memory proposals, and handoff receipts. | Use human review if role, authority, memory custody, or provenance is ambiguous. |
| L6 | Meta-Coordinator / orchestrates teams and long-running cross-domain memory | Routes evidence across teams and memory systems under explicit boundaries. It is not deployed proof of consciousness or unrestricted autonomy. | Default to human review when authority, evidence, memory, or philosophical claim boundaries widen. |
Required packet expectations
- Startup packets name read order, domain boundary, no-op triggers, and validation targets.
- Suspension packets preserve state, checks, blockers, and next safe step.
- Receiver briefs name read order, first acceptable response, domain boundaries, and escalation triggers.
- Ability profiles map LocalEndpoint-style capability surfaces without granting execution authority.
- Carcinus continuity markers remain external identity evidence until separately verified.
- Teleodynamic claim-boundary notes stop UAIX pages from widening standards records into philosophical proof.