Permission systems answer "may I?" Capability answers "can I?" — whether the substrate structurally supports the action. Capability envelope. Temporal executability forecasting. Joint condition combining capability, time, and uncertainty.
Every existing access control system answers a single question: "Is this agent authorized to perform this action?" Authorization is a governance property — a policy decision. But authorization says nothing about whether the action can actually be performed. An agent may be authorized to write to a database that is offline. It may be permitted to invoke a service that has been deprecated. It may have the right to execute a computation that exceeds available resources.
The gap between permission and capability produces a class of failures that no governance system detects: authorized actions that fail at execution because the substrate cannot support them. The agent is not misbehaving. The governance system is not failing. The action was permitted. It simply cannot be done.
Capability awareness provides a structural assessment of what the agent can actually do given current substrate conditions. The capability envelope defines the boundary of executable actions. Temporal executability forecasting predicts how that envelope will change over time. The joint condition combines capability, time, and uncertainty into a single executability assessment: can this action be performed, at this time, with acceptable uncertainty about the outcome?
When agents assess capability before attempting execution, an entire class of runtime failures disappears. The agent does not attempt actions it cannot complete. It does not promise outcomes it cannot deliver. It does not consume resources on operations that will fail. And when capability is changing over time — as it always is in dynamic distributed environments — the agent can forecast when an action will become executable and plan accordingly.
This is the missing complement to governance: not just "may you act?" but "can you act?" — and if not now, when? The combination of capability awareness with confidence governance produces agents that execute only when they are both authorized and able, with quantified uncertainty about the outcome.
Substrate-aware execution for autonomous agents. Published and available to license.
No guarantee of issuance or scope. No rights granted by this page. Any license requires issued claims (if any) and a separate written agreement.