Semantic Substrate

Inquire

The primary namespace coordinate for the agent safety discipline.

A direct, unambiguous string that names the field responsible for ensuring autonomous agents behave reliably, transparently, and within intended boundaries.

Coordinated sets this position belongs to — the coverage it extends. Counts are the live cluster size in the graph.

Primary home

Agent cluster281

Also appears in

Safety cluster48

Architectural context

Agent · Cross-Vertical · 2 compound moats. Architectural surface: Agent. Cross-cutting: Safety.

Layer position: Cross-cutting

AgentSafety

Why this is canonical

'Agent safety' is the operative phrase used by AI safety researchers, policymakers, and standards bodies to describe the subset of AI safety concerned specifically with autonomous and agentic systems. As agents move from experimental to production, this names the governing discipline. The .ai TLD confirms the domain.

Where it fits

A few directions this coordinate opens —

Research and standards
A home for the academic, policy, or standards-body work defining what it means for agents to be safe — evaluation frameworks, red-teaming protocols, and alignment benchmarks.
AI safety research organizations, standards bodies, academic labs
Enterprise risk and compliance
The product or platform that operationalizes agent safety requirements for enterprises deploying agents in regulated or high-stakes environments.
Enterprise AI governance platforms, regulated-industry AI deployers, risk management vendors

Illustrative, not exhaustive — held as a transferable canonical position, open to the buyer's own use.

← More in Agent