Trace capture

When enabled, trace capture records bounded call metadata. Coverage depends on the execution path and configuration.

The table describes the episode emitter. A declared type does not prove that an event was recorded for a run.

Episode emitter capabilities
EventEmitter supportFieldsExecution path
agent_turnImplementedAgent identity, timing, statusEpisode trace emitter
model_callImplementedModel identity, timing, token counts when available, hashesEpisode model calls
judge_callImplementedJudge identity, timing, token counts when available, hashesEpisode judge calls
sandbox_actionImplementedTool name, argument hash, bounded outcome, timingEpisode executor sandbox actions
errorImplementedError code, sanitized message, timingEpisode trace emitter
tool_callNot emittedNo generic tool event emitted by the episode emitterDeclared type
optimizer_stepNot emittedNo optimizer event emitted by the episode emitterDeclared type
callbackNot emittedNo callback event emitted by the episode emitterDeclared type

Capture conditions

Capture requires TRACE_CAPTURE_ENABLED=true and an allowed run type in TRACE_CAPTURE_RUN_TYPES. The run-type default is none.

These defaults do not establish production availability. Per-run authorization and execution policy can also prevent capture.

Limits and export

The episode emitter records hashes instead of prompt text and tool arguments. It does not record hidden model reasoning.

Sandbox actions contain bounded outcome metadata. Their nested tool metadata is not a generic tool event.

Capture is best effort. Missing events, dropped events, and write failures can prevent complete reconstruction. Traces are not tamper-proof audit records.

The trace panel provides an authorized JSON export. Export availability depends on access to the episode and its trace.

Operational retention and deletion guarantees are not yet verified here. A stored retention setting alone does not prove deletion.

Read result and trace states · Trust information