Knowledge · AI Capabilities

    Registered Action Execution

    Registered actions give business AI a fixed, validated set of operations it can perform — the difference between bounded execution and open-ended access.

    What is registered action execution?

    Registered action execution means the AI can only perform operations that exist in a server-side registry, each with declared inputs, permissions and side effects. The model chooses which registered action to call and with what arguments; it cannot invent an operation or reach the database directly.

    Key takeaways

    • The registry, not the model, defines what is possible.
    • Each action declares inputs, required role, side effects and confirmation class.
    • Validation happens server-side; a well-worded prompt cannot widen scope.
    • Unknown or unregistered action names must fail closed.
    • Appearing in a catalog is discovery, not authorization to run.

    Why open database access is the wrong design

    Giving a model direct query or write access to a production database looks flexible and is unmanageable. Every prompt becomes a potential schema change, permissions collapse into whatever the connection user can do, and nothing is reviewable afterwards.

    A registry inverts that. The surface area is enumerable, reviewable and testable, and it changes only when an engineer ships a change.

    What a single registered action declares

    • A stable identifier such as leads.create or appointments.book.
    • A typed input schema, validated before any work happens.
    • The role or capability required to invoke it.
    • Its side effects: internal write, external send, financial, destructive.
    • Whether it requires a confirmation step before it executes.

    The gap between the catalog and reality

    A capability manifest lists what a system advertises. That list is genuinely useful — it lets other software discover what exists — but it proves nothing on its own.

    An entry is only a current capability once there is a handler behind it, a permission mapping for it, and a real execution that produced a persisted record. Anything short of that is a declared capability and should be described that way.

    Failing closed

    • An unrecognized action identifier is rejected, not approximated.
    • A missing required argument stops the call rather than guessing a value.
    • A permission check failure returns a refusal that is logged.
    • An entitlement limit blocks execution and explains why.

    Where URBLD fits

    URBLD keeps a server-side action registry that every Jarvis and MCP call resolves against. A parity check runs against the catalog so an advertised action without a handler or a permission mapping is treated as drift and fails the build rather than silently reaching production.

    FAQ

    Frequently Asked Questions

    Straight answers about how URBLD runs the business end-to-end.

    More in AI Capabilities

    How to tell a chatbot from an assistant from operational AI.

    Browse AI Capabilities
    Share this page