Results, reuse, and next steps

Read the operational claim

Focus on what the final response says happened, not only on how confident or polished it sounds.

• Found means Elyxis discovered a candidate.
• Created means a resource was produced, but it may still need configuration, build, or activation.
• Ready means the required lifecycle state was reached.
• Executed means an invocation ran.
• Completed means the invocation returned a result.

The exact states available depend on the capability and resource involved.

Look for verified output

A successful tool or capability result should be grounded in server-side execution evidence. The final response may summarize that evidence, but it should not claim a mutation or returned value that no execution produced.

Suggested actions are optional next requests. They are not completed work.

Understand incomplete outcomes

An incomplete result can still be useful. The Assistant may identify the exact missing configuration, lifecycle step, permission, or input required to continue.

Do not repeatedly invoke a capability that is reported as not executable. Satisfy the stated requirement first, then ask Elyxis to verify readiness and run it.

Reuse a capability

Your request defines the task and its parameters. Once created, the capability can perform the same task with different parameter values, producing results based on each request. Reusing a capability does not mean repeating the original parameter values or returning the original result.

When the result identifies a reusable capability, refer to it by the displayed name or stable identifier in the same session. If you start a new session, provide enough context for Elyxis to resolve it again.

Include the parameter values you want to use in your follow-up request. They must fit the parameters the capability accepts. If you need to change what the capability does or which parameters it accepts, ask Elyxis to update it rather than treating that as another run.

Use /context to inspect retained capability, app, or environment selections. Use /history to review conversational context.

Verify after side effects

After creating or updating a resource, a follow-up read is the clearest verification. Ask the Assistant to show the resource or report its current lifecycle state before relying on it in another workflow.