Agent Executions

Inspect and stop individual agent executions.

📄️List the executions of an agent

`fields=full` opts into the heavy message payload; `fields=vars` adds the final agent state variables without the heavy message content; anything else keeps the compact projection. `statuses` is a comma-separated list.

📄️Get an execution of an agent

Returns the execution detail. Executions that already finished or stopped are served from a cache, running ones are always read fresh, so this endpoint can be polled to track a live execution. `applicationId` is a required query parameter. Use `POST /api/v1/agents/{agentId}/executions/{executionId}/stop/` to cancel a running execution.

📄️Resolve the context sources of an agent execution

Accepts a JSON body `{ applicationId, messageIds?, knowledgeIds?, documentIds? }`. Ids of the wrong type are silently dropped; when no valid id remains the empty payload is returned without touching the database.

📄️Stop a running execution of an agent

Signals the worker running the execution to abort the in-flight run, then flips the persisted status from `RUNNING` to `STOPPED`. This also clears orphaned executions stuck in `RUNNING` after their worker died. Idempotent: stopping an execution that already finished or stopped returns `stopped: false` with `alreadyTerminal: true`. `applicationId` is a required query parameter.