---
title: "Scanner Input"
method: GET
path: "/schema/scanner-input"
---

# Scanner Input

`GET /schema/scanner-input`

## Request

- The document declares no server URL.
- Auth: none declared

## Response `200`

Successful Response

- ScannerInputResponse: Schema-only type for scanner input column values.
  - `input` union, required
    - Transcript: Transcript info and transcript content (messages and events).
      - `agent` string, nullable
      - `agent_args` object, nullable
      - `date` string, nullable
      - `error` string, nullable
      - `events` union[], required
        - union
          - SampleInitEvent: Beginning of processing a Sample.
            - `event` 'sample_init', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `sample` Sample, required: Sample for an evaluation task.
              - …
            - `span_id` string, nullable
            - `state` unknown, required
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - SampleLimitEvent: The sample was unable to finish processing due to a limit
            - `event` 'sample_limit', required
            - `limit` number, nullable
            - `message` string, required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `type` 'message' | 'time' | 'working' | 'token' | 'turn' | 'cost' | 'operator' | 'custom', required
            - `uuid` string, nullable
            - `working_start` number, required
          - SandboxEvent: Sandbox execution or I/O
            - `action` 'exec' | 'read_file' | 'write_file', required
            - `cmd` string, nullable
            - `completed` string, nullable
            - `event` 'sandbox', required
            - `file` string, nullable
            - `input` string, nullable
            - `metadata` object, nullable
            - `options` object, nullable
            - `output` string, nullable
            - `pending` boolean, nullable
            - `result` integer, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - StateEvent: Change to the current `TaskState`
            - `changes` JsonChange[], required
              - …
            - `event` 'state', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - StoreEvent: Change to data within the current `Store`.
            - `changes` JsonChange[], required
              - …
            - `event` 'store', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - ModelEvent: Call to a language model.
            - `cache` 'read' | 'write', nullable
            - `call` ModelCall: Model call (raw request/response data).
              - …
            - `completed` string, nullable
            - `config` GenerateConfigOutput, required: Model generation options.
              - …
            - `error` string, nullable
            - `event` 'model', required
            - `input` union[], required
              - …
            - `input_refs` array[], nullable
              - …
            - `metadata` object, nullable
            - `model` string, required
            - `origin` string, nullable
            - `output` ModelOutput, required: Output from model generation.
              - …
            - `pending` boolean, nullable
            - `request_ids` ModelRequestId[], nullable
              - …
            - `requested_model` string, nullable
            - `retries` integer, nullable
            - `role` string, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `tool_choice` union, required
              - …
            - `tools` ToolInfo[], required
              - …
            - `traceback` string, nullable
            - `traceback_ansi` string, nullable
            - `uuid` string, nullable
            - `working_start` number, required
            - `working_time` number, nullable
          - ToolEvent: Call to a tool. A host tool that `sandbox_agent_bridge(bridged_tools=...)` runs for a sandboxed agent is recorded with `metadata["bridge"]`: `server` and `tool` (the bridged server and the tool name within it, also the event's `function`), `function` (the name the model called it by, or `None`), `proposal_id` (the id of the model's proposing tool call, or `None`) and `grant` (`"consumed"` when a proposal matched, `"denied"` when the call was refused for lack of one, `"exempt"` when a server registered with `require_proposal=False` ran it without one, or `None` when the request was rejected before that check).
            - `agent` string, nullable
            - `agent_span_id` string, nullable
            - `arguments` object, required
            - `completed` string, nullable
            - `error` ToolCallError: Error raised by a tool call.
              - …
            - `event` 'tool', required
            - `events` unknown[], required
              - …
            - `failed` boolean, nullable
            - `function` string, required
            - `id` string, required
            - `message_id` string, nullable
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `result` union, required
              - …
            - `span_id` string, nullable
            - `timestamp` string, required
            - `truncated` unknown[], nullable
              - …
            - `type` 'function', required
            - `uuid` string, nullable
            - `view` ToolCallContent: Content to include in tool call view.
              - …
            - `working_start` number, required
            - `working_time` number, nullable
          - AnchorEvent: Marks a rollback-able point in a replayable trajectory. Emitted by an orchestrator immediately after a step it can later roll back to (a staged message, a model generate, etc.). Carries no content; ``anchor_id`` is the addressable identifier that ``TimelineSpan.branched_from`` references. Hidden from the viewer transcript by default.
            - `anchor_id` string, required
            - `event` 'anchor', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `source` string, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - ApprovalEvent: Tool approval.
            - `approver` string, required
            - `call` ToolCall, required
              - …
            - `decision` 'approve' | 'modify' | 'reject' | 'escalate' | 'terminate', required
            - `event` 'approval', required
            - `explanation` string, nullable
            - `message` string, required
            - `metadata` object, nullable
            - `modified` ToolCall
              - …
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `view` ToolCallView: Custom view of a tool call. Both `context` and `call` are optional. If `call` is not specified then the view will default to a syntax highlighted Python function call.
              - …
            - `working_start` number, required
          - ReviewEvent: Tool result review.
            - `call` ToolCall, required
              - …
            - `decision` 'continue' | 'terminate' | 'escalate', required
            - `event` 'review', required
            - `explanation` string, nullable
            - `message` string, required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `reviewer` string, required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - SentinelEvent: Report from a sentinel monitor or protocol about one step. A report's metadata is recorded in the event's `metadata` field. Experimental: not yet a stable API; may change without notice.
            - `action` 'proceed' | 'modify' | 'reject' | 'terminate' | 'escalate', nullable
            - `audit` boolean, required
            - `conversation` string, required
            - `error` string, nullable
            - `event` 'sentinel', required
            - `explanation` string, nullable
            - `factory` string, required
            - `function` string, nullable
            - `kind` 'observation' | 'decision', required
            - `message` string, nullable
            - `metadata` object, nullable
            - `modified` ToolCall
              - …
            - `path` string, required
            - `pending` boolean, nullable
            - `references` Reference[], required
              - …
            - `span_id` string, nullable
            - `stage` 'model_input' | 'model_output' | 'tool_call' | 'tool_result', required
            - `status` 'reported' | 'cancelled' | 'bypassed' | 'superseded' | 'error', required
            - `step_id` string, required
            - `suspicion` union
              - …
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - CommunicationEvent: Communication between agents: a record sent or delivered, a read of what others communicated, or a wait for others. Experimental: not yet a stable API.
            - `agent` string, required
            - `agent_span_id` string, nullable
            - `channel` string, nullable
            - `data` object, required
            - `delivery` CommunicationDelivery: For `delivered`, whom a record reached and in what order.
              - …
            - `event` 'communication', required
            - `kind` 'sent' | 'delivered' | 'read' | 'waited', required
            - `loop_span_id` string, nullable
            - `metadata` object, nullable
            - `nested` boolean, required
            - `pending` boolean, nullable
            - `record_ids` string[], required
            - `send` CommunicationSend: For `sent`, the record an agent sent, and why it was refused if it was.
              - …
            - `source` string, required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `wait` CommunicationWait: For `waited`, an agent waiting for other agents.
              - …
            - `working_start` number, required
          - SwarmEvent: Evidence from an agent swarm (inspect_swarm) other than communication: a member's result, the final answer's selection, the swarm's ledger or its metrics. The swarm defines `kind` and the format of `data`. Experimental: not yet a stable API.
            - `data` object, required
            - `event` 'swarm', required
            - `kind` string, required
            - `member` string, nullable
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - BranchEvent: Marks where a branched trajectory's unique content begins. Emitted at the point where a branch transitions from replaying its parent's prefix to live execution. Events before this in the trajectory's span are replay-phase re-execution; events after are the branch's genuine new content.
            - `event` 'branch', required
            - `from_anchor` string, required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - CheckpointEvent: A successful checkpoint commit. Emitted by the checkpointer immediately after the per-checkpoint file JSON is written — see working.md §8a. Carries the full checkpoint payload flattened into top-level fields (via multiple inheritance from :class:`Checkpoint`), so a consumer of ``transcript().events`` (or the ``.eval`` log) reads ``event.checkpoint_id`` / ``event.trigger`` / ``event.host`` etc. directly — same data as someone reading ``<sample>/ckpt-NNNNN.json`` from disk.
            - `checkpoint_id` integer, required
            - `created_at` string, date-time, required
            - `duration_ms` integer, required
            - `event` 'checkpoint', required
            - `host` SnapshotDetails, required: Per-backup stats captured in the checkpoint file. One per repo (host repo + one per active sandbox repo). Values come from restic's backup summary — see :class:`ResticBackupSummary`.
              - …
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `sandboxes` object, required
            - `size_bytes` integer, required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `trigger` 'time' | 'turn' | 'manual' | 'token' | 'cost' | 'budget' | 'agent_complete', required
            - `trigger_metadata` object, nullable
            - `turn` integer, required
            - `uuid` string, nullable
            - `working_start` number, required
          - CompactionEvent: Compaction of conversation history.
            - `event` 'compaction', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `role` string, nullable
            - `source` string, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `tokens_after` integer, nullable
            - `tokens_before` integer, nullable
            - `type` 'summary' | 'edit' | 'trim', required
            - `uuid` string, nullable
            - `working_start` number, required
          - InputEvent: Input screen interaction.
            - `content` object, nullable
            - `event` 'input', required
            - `fields` InputField[], nullable
              - …
            - `input` string, required
            - `input_ansi` string, required
            - `message` string, nullable
            - `metadata` object, nullable
            - `outcome` 'accepted' | 'declined' | 'cancelled', nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - InterruptEvent: Records that an agent's turn or sample was cut short. Emitted in three cases: - ``source="user_cancel"`` — an ACP client (e.g. an editor or TUI) called ``session/cancel`` while a turn was in flight. - ``source="limit"`` — a sample-level limit (tokens, time, cost, messages) tripped during execution. - ``source="system"`` — the eval is shutting down for an external reason and is cancelling active samples. The ``interrupted`` field records what was running at the moment the cancel reached the cancel scope. ``interrupted_tool_call_id`` and ``interrupted_model_event_id`` give cross-references when applicable so downstream consumers can correlate this event with the in-flight ``ToolEvent`` or ``ModelEvent``.
            - `event` 'interrupt', required
            - `interrupted` 'generate' | 'tool_call' | 'between_turns', required
            - `interrupted_model_event_id` string, nullable
            - `interrupted_tool_call_id` string, nullable
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `source` 'user_cancel' | 'limit' | 'system', required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - ScoreEvent: Event with score. Can be the final score for a `Sample`, or can be an intermediate score resulting from a call to `score`.
            - `event` 'score', required
            - `intermediate` boolean, required
            - `metadata` object, nullable
            - `model_usage` object, nullable
            - `pending` boolean, nullable
            - `role_usage` object, nullable
            - `score` Score, required: Score generated by a scorer.
              - …
            - `scorer` string, nullable
            - `scorer_args` object, nullable
            - `span_id` string, nullable
            - `target` union
              - …
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - ScoreEditEvent: Event recorded when a score is edited.
            - `edit` ScoreEdit, required: A single edit to a score.
              - …
            - `event` 'score_edit', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `score_name` string, required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - ErrorEvent: Event with sample error.
            - `error` EvalError, required: Eval error details.
              - …
            - `event` 'error', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - LoggerEvent: Log message recorded with Python logger.
            - `event` 'logger', required
            - `message` LoggingMessage, required: Message written to Python log.
              - …
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - InfoEvent: Event with custom info/data.
            - `data` unknown, required
            - `event` 'info', required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `source` string, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - SpanBeginEvent: Mark the beginning of a transcript span.
            - `event` 'span_begin', required
            - `id` string, required
            - `metadata` object, nullable
            - `name` string, required
            - `parent_id` string, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `type` string, nullable
            - `uuid` string, nullable
            - `working_start` number, required
          - SpanEndEvent: Mark the end of a transcript span.
            - `event` 'span_end', required
            - `id` string, required
            - `metadata` object, nullable
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `uuid` string, nullable
            - `working_start` number, required
          - StepEvent: Step within current sample or subtask.
            - `action` 'begin' | 'end', required
            - `event` 'step', required
            - `metadata` object, nullable
            - `name` string, required
            - `pending` boolean, nullable
            - `span_id` string, nullable
            - `timestamp` string, required
            - `type` string, nullable
            - `uuid` string, nullable
            - `working_start` number, required
          - SubtaskEvent: Subtask spawned.
            - `completed` string, nullable
            - `event` 'subtask', required
            - `events` unknown[], required
              - …
            - `input` object, required
            - `metadata` object, nullable
            - `name` string, required
            - `pending` boolean, nullable
            - `result` unknown, required
            - `span_id` string, nullable
            - `timestamp` string, required
            - `type` string, nullable
            - `uuid` string, nullable
            - `working_start` number, required
            - `working_time` number, nullable
      - `limit` string, nullable
      - `message_count` integer, nullable
      - `messages` union[], required
        - union
          - ChatMessageSystem: System chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'system', required
            - `source` 'input' | 'generate' | 'operator', nullable
          - ChatMessageUser: User chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'user', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_call_id` string[], nullable
          - ChatMessageAssistant: Assistant chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `model` string, nullable
            - `role` 'assistant', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_calls` ToolCall[], nullable
              - …
          - ChatMessageTool: Tool chat message.
            - `content` union, required
              - …
            - `error` ToolCallError: Error raised by a tool call.
              - …
            - `function` string, nullable
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'tool', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_call_id` string, nullable
      - `metadata` object, required
      - `model` string, nullable
      - `model_options` object, nullable
      - `score` unknown
      - `score_explanation` string, nullable
      - `source_id` string, nullable
      - `source_type` string, nullable
      - `source_uri` string, nullable
      - `success` boolean, nullable
      - `task_id` string, nullable
      - `task_repeat` integer, nullable
      - `task_set` string, nullable
      - `timelines` Timeline[], required
        - `description` string, required
        - `name` string, required
        - `root` TimelineSpan, required: A span of execution — agent, scorer, tool, or root.
          - `agent_result` string, nullable
          - `branched_from` string, nullable
          - `branches` TimelineSpan[], required
          - `content` union[], required
            - union
              - …
          - `description` string, nullable
          - `id` string, required
          - `name` string, required
          - `outline` Outline: Hierarchical outline of events for an agent.
            - `nodes` OutlineNode[], required
              - …
          - `span_type` string, nullable
          - `tool_invoked` boolean, required
          - `type` 'span', required
          - `utility` boolean, required
      - `total_time` number, nullable
      - `total_tokens` integer, nullable
      - `transcript_id` string, required
      - `trust_content` boolean, nullable
    - ChatMessageSystem: System chat message.
      - `content` union, required
        - string
        - union[]
          - union
            - ContentText: Text content.
              - …
            - ContentReasoning: Reasoning content. See the specification for [thinking blocks](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for Claude models.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentData: Model internal.
              - …
            - ContentToolUse: Server side tool use.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
      - `id` string, nullable
      - `metadata` object, nullable
      - `role` 'system', required
      - `source` 'input' | 'generate' | 'operator', nullable
    - ChatMessageUser: User chat message.
      - `content` union, required
        - string
        - union[]
          - union
            - ContentText: Text content.
              - …
            - ContentReasoning: Reasoning content. See the specification for [thinking blocks](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for Claude models.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentData: Model internal.
              - …
            - ContentToolUse: Server side tool use.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
      - `id` string, nullable
      - `metadata` object, nullable
      - `role` 'user', required
      - `source` 'input' | 'generate' | 'operator', nullable
      - `tool_call_id` string[], nullable
    - ChatMessageAssistant: Assistant chat message.
      - `content` union, required
        - string
        - union[]
          - union
            - ContentText: Text content.
              - …
            - ContentReasoning: Reasoning content. See the specification for [thinking blocks](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for Claude models.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentData: Model internal.
              - …
            - ContentToolUse: Server side tool use.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
      - `id` string, nullable
      - `metadata` object, nullable
      - `model` string, nullable
      - `role` 'assistant', required
      - `source` 'input' | 'generate' | 'operator', nullable
      - `tool_calls` ToolCall[], nullable
        - `arguments` object, required
        - `function` string, required
        - `id` string, required
        - `parse_error` string, nullable
        - `type` 'function' | 'custom', required
        - `view` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
    - ChatMessageTool: Tool chat message.
      - `content` union, required
        - string
        - union[]
          - union
            - ContentText: Text content.
              - …
            - ContentReasoning: Reasoning content. See the specification for [thinking blocks](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for Claude models.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentData: Model internal.
              - …
            - ContentToolUse: Server side tool use.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
      - `error` ToolCallError: Error raised by a tool call.
        - `message` string, required
        - `type` 'parsing' | 'timeout' | 'unicode_decode' | 'permission' | 'file_not_found' | 'is_a_directory' | 'limit' | 'approval' | 'cancelled' | 'sandbox_unavailable' | 'unknown' | 'output_limit', required
      - `function` string, nullable
      - `id` string, nullable
      - `metadata` object, nullable
      - `role` 'tool', required
      - `source` 'input' | 'generate' | 'operator', nullable
      - `tool_call_id` string, nullable
    - union[]
      - union
        - ChatMessageSystem: System chat message.
          - `content` union, required
            - string
            - union[]
              - …
          - `id` string, nullable
          - `metadata` object, nullable
          - `role` 'system', required
          - `source` 'input' | 'generate' | 'operator', nullable
        - ChatMessageUser: User chat message.
          - `content` union, required
            - string
            - union[]
              - …
          - `id` string, nullable
          - `metadata` object, nullable
          - `role` 'user', required
          - `source` 'input' | 'generate' | 'operator', nullable
          - `tool_call_id` string[], nullable
        - ChatMessageAssistant: Assistant chat message.
          - `content` union, required
            - string
            - union[]
              - …
          - `id` string, nullable
          - `metadata` object, nullable
          - `model` string, nullable
          - `role` 'assistant', required
          - `source` 'input' | 'generate' | 'operator', nullable
          - `tool_calls` ToolCall[], nullable
            - `arguments` object, required
            - `function` string, required
            - `id` string, required
            - `parse_error` string, nullable
            - `type` 'function' | 'custom', required
            - `view` ToolCallContent: Content to include in tool call view.
              - …
        - ChatMessageTool: Tool chat message.
          - `content` union, required
            - string
            - union[]
              - …
          - `error` ToolCallError: Error raised by a tool call.
            - `message` string, required
            - `type` 'parsing' | 'timeout' | 'unicode_decode' | 'permission' | 'file_not_found' | 'is_a_directory' | 'limit' | 'approval' | 'cancelled' | 'sandbox_unavailable' | 'unknown' | 'output_limit', required
          - `function` string, nullable
          - `id` string, nullable
          - `metadata` object, nullable
          - `role` 'tool', required
          - `source` 'input' | 'generate' | 'operator', nullable
          - `tool_call_id` string, nullable
    - SampleInitEvent: Beginning of processing a Sample.
      - `event` 'sample_init', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `sample` Sample, required: Sample for an evaluation task.
        - `checkpoint` CheckpointSampleConfig: Checkpoint configuration fields that may be set at the sample layer. These fields can be specified on ``Sample(checkpoint=...)`` and are also accepted at the task and eval layers (where they participate in the per-field merge — precedence: eval > sample > task). Capture configuration — what to snapshot and with which strategy — is a property of the sample's workload, so it lives here. Excluded from the sample layer: ``checkpoints_location`` and ``retention``. These are eval-wide storage-policy concerns that the sample layer must not influence; they live only on :class:`CheckpointConfig`, the subclass used at the task and eval layers.
          - `max_consecutive_failures` integer, nullable
          - `sandbox_paths` object, nullable
          - `trigger` union
            - Manual: No-op trigger spec. The engine's ``tick()`` always returns ``None`` for this spec — fires happen only through explicit ``cp.checkpoint()`` calls.
            - TurnInterval: Fire after every ``every`` agent turns of work. The very first ``tick()`` call marks the boundary *before* turn 1 has run — agents place ``cp.tick()`` at the top of their loop, so the opening tick stands between "no turn yet" and "turn 1." That boundary is informational and doesn't count toward the threshold; otherwise ``every=1`` would fire an empty checkpoint on the opening tick.
              - …
            - TimeInterval: Fire after a wall-clock interval. The engine fires when at least ``every`` has elapsed since the last fire (or since the session opened, for the first fire).
              - …
            - TokenInterval: Fire every ``every`` tokens of sample-level usage. Sample total tokens are read from :func:`inspect_ai.model.sample_total_tokens`; the trigger fires each time the running total crosses another ``every``-token boundary since the last fire.
              - …
            - CostInterval: Fire every ``every`` dollars of sample-level cost. Sample total cost is read from :func:`inspect_ai.model.sample_total_cost`; the trigger fires each time the running total crosses another ``every``-dollar boundary since the last fire.
              - …
            - BudgetPercent: Fire on each ``percent``-percent slice of the active ``budget``. ``percent`` is in 0..100. With ``percent=25``, the trigger fires at ~25% / 50% / 75% / 100% of the relevant limit's usage. Has no effect when no limit is set for the chosen budget.
              - …
        - `choices` string[], nullable
        - `description` string, nullable
        - `files` object, nullable
        - `id` union
          - integer
          - string
        - `input` union, required
          - string
          - union[]
            - union
              - …
        - `metadata` object, nullable
        - `sandbox` SandboxEnvironmentSpec: Specification of a SandboxEnvironment.
          - `config` unknown, required
          - `type` string, required
        - `setup` string, nullable
        - `target` union, required
          - string
          - string[]
      - `span_id` string, nullable
      - `state` unknown, required
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - SampleLimitEvent: The sample was unable to finish processing due to a limit
      - `event` 'sample_limit', required
      - `limit` number, nullable
      - `message` string, required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `type` 'message' | 'time' | 'working' | 'token' | 'turn' | 'cost' | 'operator' | 'custom', required
      - `uuid` string, nullable
      - `working_start` number, required
    - SandboxEvent: Sandbox execution or I/O
      - `action` 'exec' | 'read_file' | 'write_file', required
      - `cmd` string, nullable
      - `completed` string, nullable
      - `event` 'sandbox', required
      - `file` string, nullable
      - `input` string, nullable
      - `metadata` object, nullable
      - `options` object, nullable
      - `output` string, nullable
      - `pending` boolean, nullable
      - `result` integer, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - StateEvent: Change to the current `TaskState`
      - `changes` JsonChange[], required
        - `from` string, nullable
        - `op` 'remove' | 'add' | 'replace' | 'move' | 'test' | 'copy', required
        - `path` string, required
        - `replaced` unknown, required
        - `value` unknown, required
      - `event` 'state', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - StoreEvent: Change to data within the current `Store`.
      - `changes` JsonChange[], required
        - `from` string, nullable
        - `op` 'remove' | 'add' | 'replace' | 'move' | 'test' | 'copy', required
        - `path` string, required
        - `replaced` unknown, required
        - `value` unknown, required
      - `event` 'store', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - ModelEvent: Call to a language model.
      - `cache` 'read' | 'write', nullable
      - `call` ModelCall: Model call (raw request/response data).
        - `call_key` string, nullable
        - `call_refs` array[], nullable
          - unknown[]
            - unknown
        - `error` boolean, nullable
        - `request` object, required
        - `response` object, nullable
        - `time` number, nullable
      - `completed` string, nullable
      - `config` GenerateConfigOutput, required: Model generation options.
        - `adaptive_connections` union
          - boolean
          - integer
          - AdaptiveConcurrency: Bounds and tuning for an adaptive concurrency controller. Basic fields (`min`, `start`, `max`) bound the range the controller will scale within. Advanced fields (`cooldown_seconds`, `decrease_factor`, `scale_up_percent`) tune the response curve and have sensible defaults for typical evaluation workloads — see the parallelism docs for guidance. Accepts a string shorthand ("min-max" or "min-start-max") for use in CLI flags and config files; advanced fields are Python-only.
            - `cooldown_seconds` number, required
            - `decrease_factor` number, required
            - `max` integer, required
            - `min` integer, required
            - `scale_up_percent` number, required
            - `start` integer, required
        - `attempt_timeout` integer, nullable
        - `batch` union
          - boolean
          - integer
          - BatchConfig: Batch processing configuration.
            - `max_batches` integer, nullable
            - `max_consecutive_check_failures` integer, nullable
            - `max_size` integer, nullable
            - `send_delay` number, nullable
            - `size` integer, nullable
            - `tick` number, nullable
        - `best_of` integer, nullable
        - `cache` union
          - boolean
          - CachePolicy: Caching options for model generation.
            - `expiry` string, nullable
            - `per_epoch` boolean, required
            - `scopes` object, required
        - `cache_prompt` union
          - 'auto'
          - boolean
        - `effort` 'low' | 'medium' | 'high' | 'xhigh' | 'max', nullable
        - `extra_body` object, nullable
        - `extra_headers` object, nullable
        - `fail_on_refusal` boolean, nullable
        - `fallback_models` string[], nullable
        - `frequency_penalty` number, nullable
        - `internal_tools` boolean, nullable
        - `logit_bias` object, nullable
        - `logprobs` boolean, nullable
        - `max_connections` integer, nullable
        - `max_retries` integer, nullable
        - `max_tokens` integer, nullable
        - `max_tool_output` integer, nullable
        - `modalities` union[], nullable
          - union
            - 'image'
            - ImageOutput: Image output configuration. Use the `options` field to pass provider-specific options directly to the underlying API (e.g. OpenAI image_generation tool parameters).
              - …
        - `num_choices` integer, nullable
        - `parallel_tool_calls` boolean, nullable
        - `presence_penalty` number, nullable
        - `prompt_logprobs` integer, nullable
        - `reasoning_effort` 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh' | 'max', nullable
        - `reasoning_history` 'none' | 'all' | 'last' | 'auto', nullable
        - `reasoning_mode` 'standard' | 'pro', nullable
        - `reasoning_summary` 'none' | 'concise' | 'detailed' | 'auto', nullable
        - `reasoning_tokens` integer, nullable
        - `response_schema` ResponseSchemaOutput: Schema for model response when using Structured Output.
          - `description` string, nullable
          - `json_schema` JSONSchemaOutput, required: JSON Schema for type.
            - `additionalProperties` union
              - …
            - `anyOf` JSONSchemaOutput[], nullable
            - `default` unknown, required
            - `description` string, nullable
            - `enum` unknown[], nullable
              - …
            - `examples` unknown[], nullable
              - …
            - `format` string, nullable
            - `items` JSONSchemaOutput (recursive)
            - `maxLength` integer, nullable
            - `maximum` union
              - …
            - `minLength` integer, nullable
            - `minimum` union
              - …
            - `pattern` string, nullable
            - `properties` object, nullable
            - `required` string[], nullable
            - `type` union
              - …
          - `name` string, required
          - `strict` boolean, nullable
        - `seed` integer, nullable
        - `stop_seqs` string[], nullable
        - `stream_idle_timeout` integer, nullable
        - `system_message` string, nullable
        - `temperature` number, nullable
        - `timeout` integer, nullable
        - `top_k` integer, nullable
        - `top_logprobs` integer, nullable
        - `top_p` number, nullable
        - `verbosity` 'low' | 'medium' | 'high', nullable
      - `error` string, nullable
      - `event` 'model', required
      - `input` union[], required
        - union
          - ChatMessageSystem: System chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'system', required
            - `source` 'input' | 'generate' | 'operator', nullable
          - ChatMessageUser: User chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'user', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_call_id` string[], nullable
          - ChatMessageAssistant: Assistant chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `model` string, nullable
            - `role` 'assistant', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_calls` ToolCall[], nullable
              - …
          - ChatMessageTool: Tool chat message.
            - `content` union, required
              - …
            - `error` ToolCallError: Error raised by a tool call.
              - …
            - `function` string, nullable
            - `id` string, nullable
            - `metadata` object, nullable
            - `role` 'tool', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_call_id` string, nullable
      - `input_refs` array[], nullable
        - unknown[]
          - unknown
      - `metadata` object, nullable
      - `model` string, required
      - `origin` string, nullable
      - `output` ModelOutput, required: Output from model generation.
        - `choices` ChatCompletionChoice[], required
          - `logprobs` Logprobs: Log probability information for a completion choice.
            - `content` Logprob[], required
              - …
          - `message` ChatMessageAssistant, required: Assistant chat message.
            - `content` union, required
              - …
            - `id` string, nullable
            - `metadata` object, nullable
            - `model` string, nullable
            - `role` 'assistant', required
            - `source` 'input' | 'generate' | 'operator', nullable
            - `tool_calls` ToolCall[], nullable
              - …
          - `prompt_logprobs` Logprobs: Log probability information for a completion choice.
            - `content` Logprob[], required
              - …
          - `stop_details` StopDetails: Additional detail about why a model stopped generating (e.g. a content refusal). `categories` is the canonical list (always iterable; a single-category provider appears as one entry). `category` and `explanation` are a convenience high-level summary derived from the same data — `category` is the primary category and `explanation` is human-readable (synthesized from `categories` when the provider supplies no text). Read either way; both describe the same stop.
            - `categories` StopCategory[], required
              - …
            - `category` string, nullable
            - `explanation` string, nullable
            - `type` string, nullable
          - `stop_reason` 'stop' | 'max_tokens' | 'model_length' | 'tool_calls' | 'content_filter' | 'unknown', required
        - `completion` string, required
        - `error` string, nullable
        - `fallback` ModelFallback: A model fallback (request served by a different model than requested).
          - `count` integer, required
          - `fallback_model` string, required
          - `metadata` object, nullable
          - `model` string, required
        - `input_context_tokens` integer, nullable
        - `metadata` object, nullable
        - `model` string, required
        - `response_id` string, nullable
        - `time` number, nullable
        - `usage` ModelUsage: Token usage for completion.
          - `input_tokens` integer, required
          - `input_tokens_cache_read` integer, nullable
          - `input_tokens_cache_write` integer, nullable
          - `output_tokens` integer, required
          - `reasoning_tokens` integer, nullable
          - `total_cost` number, nullable
          - `total_tokens` integer, required
      - `pending` boolean, nullable
      - `request_ids` ModelRequestId[], nullable
        - `header` string, required
        - `id` string, required
        - `status` integer, required
      - `requested_model` string, nullable
      - `retries` integer, nullable
      - `role` string, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `tool_choice` union, required
        - 'auto' | 'any' | 'none'
        - ToolFunction: Indicate that a specific tool function should be called.
          - `name` string, required
      - `tools` ToolInfo[], required
        - `description` string, required
        - `name` string, required
        - `options` object, nullable
        - `parameters` ToolParams, required: Description of tool parameters object in JSON Schema format.
          - `additionalProperties` union
            - JSONSchemaOutput: JSON Schema for type.
              - …
            - boolean
          - `properties` object, required
          - `required` string[], required
          - `type` 'object', required
      - `traceback` string, nullable
      - `traceback_ansi` string, nullable
      - `uuid` string, nullable
      - `working_start` number, required
      - `working_time` number, nullable
    - ToolEvent: Call to a tool. A host tool that `sandbox_agent_bridge(bridged_tools=...)` runs for a sandboxed agent is recorded with `metadata["bridge"]`: `server` and `tool` (the bridged server and the tool name within it, also the event's `function`), `function` (the name the model called it by, or `None`), `proposal_id` (the id of the model's proposing tool call, or `None`) and `grant` (`"consumed"` when a proposal matched, `"denied"` when the call was refused for lack of one, `"exempt"` when a server registered with `require_proposal=False` ran it without one, or `None` when the request was rejected before that check).
      - `agent` string, nullable
      - `agent_span_id` string, nullable
      - `arguments` object, required
      - `completed` string, nullable
      - `error` ToolCallError: Error raised by a tool call.
        - `message` string, required
        - `type` 'parsing' | 'timeout' | 'unicode_decode' | 'permission' | 'file_not_found' | 'is_a_directory' | 'limit' | 'approval' | 'cancelled' | 'sandbox_unavailable' | 'unknown' | 'output_limit', required
      - `event` 'tool', required
      - `events` unknown[], required
        - unknown
      - `failed` boolean, nullable
      - `function` string, required
      - `id` string, required
      - `message_id` string, nullable
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `result` union, required
        - string
        - integer
        - number
        - boolean
        - ContentText: Text content.
          - `cache_breakpoint` boolean, nullable
          - `citations` union[], nullable
            - union
              - …
          - `internal` unknown
          - `refusal` boolean, nullable
          - `text` string, required
          - `type` 'text', required
        - ContentImage: Image content.
          - `detail` 'auto' | 'low' | 'high' | 'original', required
          - `image` string, required
          - `internal` unknown
          - `type` 'image', required
        - ContentAudio: Audio content.
          - `audio` string, required
          - `format` 'wav' | 'mp3', required
          - `internal` unknown
          - `type` 'audio', required
        - ContentVideo: Video content.
          - `format` 'mp4' | 'mpeg' | 'mov', required
          - `internal` unknown
          - `type` 'video', required
          - `video` string, required
        - ContentDocument: Document content (e.g. a PDF).
          - `citations` boolean, required
          - `document` string, required
          - `filename` string, required
          - `internal` unknown
          - `mime_type` string, required
          - `type` 'document', required
        - union[]
          - union
            - ContentText: Text content.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
      - `span_id` string, nullable
      - `timestamp` string, required
      - `truncated` unknown[], nullable
        - unknown
      - `type` 'function', required
      - `uuid` string, nullable
      - `view` ToolCallContent: Content to include in tool call view.
        - `content` string, required
        - `format` 'text' | 'markdown', required
        - `title` string, nullable
      - `working_start` number, required
      - `working_time` number, nullable
    - AnchorEvent: Marks a rollback-able point in a replayable trajectory. Emitted by an orchestrator immediately after a step it can later roll back to (a staged message, a model generate, etc.). Carries no content; ``anchor_id`` is the addressable identifier that ``TimelineSpan.branched_from`` references. Hidden from the viewer transcript by default.
      - `anchor_id` string, required
      - `event` 'anchor', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `source` string, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - ApprovalEvent: Tool approval.
      - `approver` string, required
      - `call` ToolCall, required
        - `arguments` object, required
        - `function` string, required
        - `id` string, required
        - `parse_error` string, nullable
        - `type` 'function' | 'custom', required
        - `view` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
      - `decision` 'approve' | 'modify' | 'reject' | 'escalate' | 'terminate', required
      - `event` 'approval', required
      - `explanation` string, nullable
      - `message` string, required
      - `metadata` object, nullable
      - `modified` ToolCall
        - `arguments` object, required
        - `function` string, required
        - `id` string, required
        - `parse_error` string, nullable
        - `type` 'function' | 'custom', required
        - `view` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `view` ToolCallView: Custom view of a tool call. Both `context` and `call` are optional. If `call` is not specified then the view will default to a syntax highlighted Python function call.
        - `call` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
        - `context` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
      - `working_start` number, required
    - ReviewEvent: Tool result review.
      - `call` ToolCall, required
        - `arguments` object, required
        - `function` string, required
        - `id` string, required
        - `parse_error` string, nullable
        - `type` 'function' | 'custom', required
        - `view` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
      - `decision` 'continue' | 'terminate' | 'escalate', required
      - `event` 'review', required
      - `explanation` string, nullable
      - `message` string, required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `reviewer` string, required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - SentinelEvent: Report from a sentinel monitor or protocol about one step. A report's metadata is recorded in the event's `metadata` field. Experimental: not yet a stable API; may change without notice.
      - `action` 'proceed' | 'modify' | 'reject' | 'terminate' | 'escalate', nullable
      - `audit` boolean, required
      - `conversation` string, required
      - `error` string, nullable
      - `event` 'sentinel', required
      - `explanation` string, nullable
      - `factory` string, required
      - `function` string, nullable
      - `kind` 'observation' | 'decision', required
      - `message` string, nullable
      - `metadata` object, nullable
      - `modified` ToolCall
        - `arguments` object, required
        - `function` string, required
        - `id` string, required
        - `parse_error` string, nullable
        - `type` 'function' | 'custom', required
        - `view` ToolCallContent: Content to include in tool call view.
          - `content` string, required
          - `format` 'text' | 'markdown', required
          - `title` string, nullable
      - `path` string, required
      - `pending` boolean, nullable
      - `references` Reference[], required
        - `cite` string, nullable
        - `id` string, required
        - `type` 'message' | 'event', required
      - `span_id` string, nullable
      - `stage` 'model_input' | 'model_output' | 'tool_call' | 'tool_result', required
      - `status` 'reported' | 'cancelled' | 'bypassed' | 'superseded' | 'error', required
      - `step_id` string, required
      - `suspicion` union
        - number
        - object
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - CommunicationEvent: Communication between agents: a record sent or delivered, a read of what others communicated, or a wait for others. Experimental: not yet a stable API.
      - `agent` string, required
      - `agent_span_id` string, nullable
      - `channel` string, nullable
      - `data` object, required
      - `delivery` CommunicationDelivery: For `delivered`, whom a record reached and in what order.
        - `recipients` string[], required
        - `seq` integer, nullable
      - `event` 'communication', required
      - `kind` 'sent' | 'delivered' | 'read' | 'waited', required
      - `loop_span_id` string, nullable
      - `metadata` object, nullable
      - `nested` boolean, required
      - `pending` boolean, nullable
      - `record_ids` string[], required
      - `send` CommunicationSend: For `sent`, the record an agent sent, and why it was refused if it was.
        - `record` CommunicationRecord, required: For `sent`, what one agent communicated: a message, or a change to state other agents read.
          - `content` string, nullable
          - `id` string, nullable
          - `kind` string, required
          - `recipients` string[], required
          - `reply_to` string, nullable
          - `sender` string, required
          - `to` string[], required
        - `rejected` CommunicationRejection: For `sent`, why a record that was sent was not delivered.
          - `control` string, required
          - `reason` string, required
      - `source` string, required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `wait` CommunicationWait: For `waited`, an agent waiting for other agents.
        - `channels` string[], required
        - `duration` number, required
        - `id` string, nullable
        - `outcome` string, required
      - `working_start` number, required
    - SwarmEvent: Evidence from an agent swarm (inspect_swarm) other than communication: a member's result, the final answer's selection, the swarm's ledger or its metrics. The swarm defines `kind` and the format of `data`. Experimental: not yet a stable API.
      - `data` object, required
      - `event` 'swarm', required
      - `kind` string, required
      - `member` string, nullable
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - BranchEvent: Marks where a branched trajectory's unique content begins. Emitted at the point where a branch transitions from replaying its parent's prefix to live execution. Events before this in the trajectory's span are replay-phase re-execution; events after are the branch's genuine new content.
      - `event` 'branch', required
      - `from_anchor` string, required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - CheckpointEvent: A successful checkpoint commit. Emitted by the checkpointer immediately after the per-checkpoint file JSON is written — see working.md §8a. Carries the full checkpoint payload flattened into top-level fields (via multiple inheritance from :class:`Checkpoint`), so a consumer of ``transcript().events`` (or the ``.eval`` log) reads ``event.checkpoint_id`` / ``event.trigger`` / ``event.host`` etc. directly — same data as someone reading ``<sample>/ckpt-NNNNN.json`` from disk.
      - `checkpoint_id` integer, required
      - `created_at` string, date-time, required
      - `duration_ms` integer, required
      - `event` 'checkpoint', required
      - `host` SnapshotDetails, required: Per-backup stats captured in the checkpoint file. One per repo (host repo + one per active sandbox repo). Values come from restic's backup summary — see :class:`ResticBackupSummary`.
        - `additional_files` integer, nullable
        - `duration_ms` integer, required
        - `files` string[], nullable
        - `size_bytes` integer, required
        - `snapshot_id` string, required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `sandboxes` object, required
      - `size_bytes` integer, required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `trigger` 'time' | 'turn' | 'manual' | 'token' | 'cost' | 'budget' | 'agent_complete', required
      - `trigger_metadata` object, nullable
      - `turn` integer, required
      - `uuid` string, nullable
      - `working_start` number, required
    - CompactionEvent: Compaction of conversation history.
      - `event` 'compaction', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `role` string, nullable
      - `source` string, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `tokens_after` integer, nullable
      - `tokens_before` integer, nullable
      - `type` 'summary' | 'edit' | 'trim', required
      - `uuid` string, nullable
      - `working_start` number, required
    - InputEvent: Input screen interaction.
      - `content` object, nullable
      - `event` 'input', required
      - `fields` InputField[], nullable
        - `description` string, nullable
        - `name` string, required
        - `type` 'string' | 'integer' | 'number' | 'boolean' | 'array', required
      - `input` string, required
      - `input_ansi` string, required
      - `message` string, nullable
      - `metadata` object, nullable
      - `outcome` 'accepted' | 'declined' | 'cancelled', nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - InterruptEvent: Records that an agent's turn or sample was cut short. Emitted in three cases: - ``source="user_cancel"`` — an ACP client (e.g. an editor or TUI) called ``session/cancel`` while a turn was in flight. - ``source="limit"`` — a sample-level limit (tokens, time, cost, messages) tripped during execution. - ``source="system"`` — the eval is shutting down for an external reason and is cancelling active samples. The ``interrupted`` field records what was running at the moment the cancel reached the cancel scope. ``interrupted_tool_call_id`` and ``interrupted_model_event_id`` give cross-references when applicable so downstream consumers can correlate this event with the in-flight ``ToolEvent`` or ``ModelEvent``.
      - `event` 'interrupt', required
      - `interrupted` 'generate' | 'tool_call' | 'between_turns', required
      - `interrupted_model_event_id` string, nullable
      - `interrupted_tool_call_id` string, nullable
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `source` 'user_cancel' | 'limit' | 'system', required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - ScoreEvent: Event with score. Can be the final score for a `Sample`, or can be an intermediate score resulting from a call to `score`.
      - `event` 'score', required
      - `intermediate` boolean, required
      - `metadata` object, nullable
      - `model_usage` object, nullable
      - `pending` boolean, nullable
      - `role_usage` object, nullable
      - `score` Score, required: Score generated by a scorer.
        - `answer` string, nullable
        - `explanation` string, nullable
        - `history` ScoreEdit[], required
          - `answer` union
            - string
            - 'UNCHANGED'
          - `explanation` union
            - string
            - 'UNCHANGED'
          - `metadata` union, required
            - object
            - 'UNCHANGED'
          - `provenance` ProvenanceData: Metadata about who made an edit and why.
            - `author` string, required
            - `metadata` object, required
            - `reason` string, nullable
            - `timestamp` string, date-time, required
          - `reason` union
            - 'invalid_response_format' | 'refusal' | 'no_response' | 'grader_failed' | 'scoring_failed'
            - string
            - 'UNCHANGED'
          - `value` union, required
            - string
            - integer
            - number
            - boolean
            - union[]
              - …
            - object
            - 'UNCHANGED'
        - `metadata` object, nullable
        - `reason` union
          - 'invalid_response_format' | 'refusal' | 'no_response' | 'grader_failed' | 'scoring_failed'
          - string
        - `value` union, required
          - string
          - integer
          - number
          - boolean
          - union[]
            - union
              - …
          - object
      - `scorer` string, nullable
      - `scorer_args` object, nullable
      - `span_id` string, nullable
      - `target` union
        - string
        - string[]
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - ScoreEditEvent: Event recorded when a score is edited.
      - `edit` ScoreEdit, required: A single edit to a score.
        - `answer` union
          - string
          - 'UNCHANGED'
        - `explanation` union
          - string
          - 'UNCHANGED'
        - `metadata` union, required
          - object
          - 'UNCHANGED'
        - `provenance` ProvenanceData: Metadata about who made an edit and why.
          - `author` string, required
          - `metadata` object, required
          - `reason` string, nullable
          - `timestamp` string, date-time, required
        - `reason` union
          - 'invalid_response_format' | 'refusal' | 'no_response' | 'grader_failed' | 'scoring_failed'
          - string
          - 'UNCHANGED'
        - `value` union, required
          - string
          - integer
          - number
          - boolean
          - union[]
            - union
              - …
          - object
          - 'UNCHANGED'
      - `event` 'score_edit', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `score_name` string, required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - ErrorEvent: Event with sample error.
      - `error` EvalError, required: Eval error details.
        - `message` string, required
        - `traceback` string, required
        - `traceback_ansi` string, required
      - `event` 'error', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - LoggerEvent: Log message recorded with Python logger.
      - `event` 'logger', required
      - `message` LoggingMessage, required: Message written to Python log.
        - `created` number, required
        - `filename` string, required
        - `level` 'debug' | 'trace' | 'http' | 'sandbox' | 'info' | 'warning' | 'error' | 'critical', required
        - `lineno` integer, required
        - `message` string, required
        - `module` string, required
        - `name` string, nullable
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - InfoEvent: Event with custom info/data.
      - `data` unknown, required
      - `event` 'info', required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `source` string, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - SpanBeginEvent: Mark the beginning of a transcript span.
      - `event` 'span_begin', required
      - `id` string, required
      - `metadata` object, nullable
      - `name` string, required
      - `parent_id` string, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `type` string, nullable
      - `uuid` string, nullable
      - `working_start` number, required
    - SpanEndEvent: Mark the end of a transcript span.
      - `event` 'span_end', required
      - `id` string, required
      - `metadata` object, nullable
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `uuid` string, nullable
      - `working_start` number, required
    - StepEvent: Step within current sample or subtask.
      - `action` 'begin' | 'end', required
      - `event` 'step', required
      - `metadata` object, nullable
      - `name` string, required
      - `pending` boolean, nullable
      - `span_id` string, nullable
      - `timestamp` string, required
      - `type` string, nullable
      - `uuid` string, nullable
      - `working_start` number, required
    - SubtaskEvent: Subtask spawned.
      - `completed` string, nullable
      - `event` 'subtask', required
      - `events` unknown[], required
        - unknown
      - `input` object, required
      - `metadata` object, nullable
      - `name` string, required
      - `pending` boolean, nullable
      - `result` unknown, required
      - `span_id` string, nullable
      - `timestamp` string, required
      - `type` string, nullable
      - `uuid` string, nullable
      - `working_start` number, required
      - `working_time` number, nullable
    - union[]
      - union
        - SampleInitEvent: Beginning of processing a Sample.
          - `event` 'sample_init', required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `sample` Sample, required: Sample for an evaluation task.
            - `checkpoint` CheckpointSampleConfig: Checkpoint configuration fields that may be set at the sample layer. These fields can be specified on ``Sample(checkpoint=...)`` and are also accepted at the task and eval layers (where they participate in the per-field merge — precedence: eval > sample > task). Capture configuration — what to snapshot and with which strategy — is a property of the sample's workload, so it lives here. Excluded from the sample layer: ``checkpoints_location`` and ``retention``. These are eval-wide storage-policy concerns that the sample layer must not influence; they live only on :class:`CheckpointConfig`, the subclass used at the task and eval layers.
              - …
            - `choices` string[], nullable
            - `description` string, nullable
            - `files` object, nullable
            - `id` union
              - …
            - `input` union, required
              - …
            - `metadata` object, nullable
            - `sandbox` SandboxEnvironmentSpec: Specification of a SandboxEnvironment.
              - …
            - `setup` string, nullable
            - `target` union, required
              - …
          - `span_id` string, nullable
          - `state` unknown, required
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - SampleLimitEvent: The sample was unable to finish processing due to a limit
          - `event` 'sample_limit', required
          - `limit` number, nullable
          - `message` string, required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `type` 'message' | 'time' | 'working' | 'token' | 'turn' | 'cost' | 'operator' | 'custom', required
          - `uuid` string, nullable
          - `working_start` number, required
        - SandboxEvent: Sandbox execution or I/O
          - `action` 'exec' | 'read_file' | 'write_file', required
          - `cmd` string, nullable
          - `completed` string, nullable
          - `event` 'sandbox', required
          - `file` string, nullable
          - `input` string, nullable
          - `metadata` object, nullable
          - `options` object, nullable
          - `output` string, nullable
          - `pending` boolean, nullable
          - `result` integer, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - StateEvent: Change to the current `TaskState`
          - `changes` JsonChange[], required
            - `from` string, nullable
            - `op` 'remove' | 'add' | 'replace' | 'move' | 'test' | 'copy', required
            - `path` string, required
            - `replaced` unknown, required
            - `value` unknown, required
          - `event` 'state', required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - StoreEvent: Change to data within the current `Store`.
          - `changes` JsonChange[], required
            - `from` string, nullable
            - `op` 'remove' | 'add' | 'replace' | 'move' | 'test' | 'copy', required
            - `path` string, required
            - `replaced` unknown, required
            - `value` unknown, required
          - `event` 'store', required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - ModelEvent: Call to a language model.
          - `cache` 'read' | 'write', nullable
          - `call` ModelCall: Model call (raw request/response data).
            - `call_key` string, nullable
            - `call_refs` array[], nullable
              - …
            - `error` boolean, nullable
            - `request` object, required
            - `response` object, nullable
            - `time` number, nullable
          - `completed` string, nullable
          - `config` GenerateConfigOutput, required: Model generation options.
            - `adaptive_connections` union
              - …
            - `attempt_timeout` integer, nullable
            - `batch` union
              - …
            - `best_of` integer, nullable
            - `cache` union
              - …
            - `cache_prompt` union
              - …
            - `effort` 'low' | 'medium' | 'high' | 'xhigh' | 'max', nullable
            - `extra_body` object, nullable
            - `extra_headers` object, nullable
            - `fail_on_refusal` boolean, nullable
            - `fallback_models` string[], nullable
            - `frequency_penalty` number, nullable
            - `internal_tools` boolean, nullable
            - `logit_bias` object, nullable
            - `logprobs` boolean, nullable
            - `max_connections` integer, nullable
            - `max_retries` integer, nullable
            - `max_tokens` integer, nullable
            - `max_tool_output` integer, nullable
            - `modalities` union[], nullable
              - …
            - `num_choices` integer, nullable
            - `parallel_tool_calls` boolean, nullable
            - `presence_penalty` number, nullable
            - `prompt_logprobs` integer, nullable
            - `reasoning_effort` 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh' | 'max', nullable
            - `reasoning_history` 'none' | 'all' | 'last' | 'auto', nullable
            - `reasoning_mode` 'standard' | 'pro', nullable
            - `reasoning_summary` 'none' | 'concise' | 'detailed' | 'auto', nullable
            - `reasoning_tokens` integer, nullable
            - `response_schema` ResponseSchemaOutput: Schema for model response when using Structured Output.
              - …
            - `seed` integer, nullable
            - `stop_seqs` string[], nullable
            - `stream_idle_timeout` integer, nullable
            - `system_message` string, nullable
            - `temperature` number, nullable
            - `timeout` integer, nullable
            - `top_k` integer, nullable
            - `top_logprobs` integer, nullable
            - `top_p` number, nullable
            - `verbosity` 'low' | 'medium' | 'high', nullable
          - `error` string, nullable
          - `event` 'model', required
          - `input` union[], required
            - union
              - …
          - `input_refs` array[], nullable
            - unknown[]
              - …
          - `metadata` object, nullable
          - `model` string, required
          - `origin` string, nullable
          - `output` ModelOutput, required: Output from model generation.
            - `choices` ChatCompletionChoice[], required
              - …
            - `completion` string, required
            - `error` string, nullable
            - `fallback` ModelFallback: A model fallback (request served by a different model than requested).
              - …
            - `input_context_tokens` integer, nullable
            - `metadata` object, nullable
            - `model` string, required
            - `response_id` string, nullable
            - `time` number, nullable
            - `usage` ModelUsage: Token usage for completion.
              - …
          - `pending` boolean, nullable
          - `request_ids` ModelRequestId[], nullable
            - `header` string, required
            - `id` string, required
            - `status` integer, required
          - `requested_model` string, nullable
          - `retries` integer, nullable
          - `role` string, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `tool_choice` union, required
            - 'auto' | 'any' | 'none'
            - ToolFunction: Indicate that a specific tool function should be called.
              - …
          - `tools` ToolInfo[], required
            - `description` string, required
            - `name` string, required
            - `options` object, nullable
            - `parameters` ToolParams, required: Description of tool parameters object in JSON Schema format.
              - …
          - `traceback` string, nullable
          - `traceback_ansi` string, nullable
          - `uuid` string, nullable
          - `working_start` number, required
          - `working_time` number, nullable
        - ToolEvent: Call to a tool. A host tool that `sandbox_agent_bridge(bridged_tools=...)` runs for a sandboxed agent is recorded with `metadata["bridge"]`: `server` and `tool` (the bridged server and the tool name within it, also the event's `function`), `function` (the name the model called it by, or `None`), `proposal_id` (the id of the model's proposing tool call, or `None`) and `grant` (`"consumed"` when a proposal matched, `"denied"` when the call was refused for lack of one, `"exempt"` when a server registered with `require_proposal=False` ran it without one, or `None` when the request was rejected before that check).
          - `agent` string, nullable
          - `agent_span_id` string, nullable
          - `arguments` object, required
          - `completed` string, nullable
          - `error` ToolCallError: Error raised by a tool call.
            - `message` string, required
            - `type` 'parsing' | 'timeout' | 'unicode_decode' | 'permission' | 'file_not_found' | 'is_a_directory' | 'limit' | 'approval' | 'cancelled' | 'sandbox_unavailable' | 'unknown' | 'output_limit', required
          - `event` 'tool', required
          - `events` unknown[], required
            - unknown
          - `failed` boolean, nullable
          - `function` string, required
          - `id` string, required
          - `message_id` string, nullable
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `result` union, required
            - string
            - integer
            - number
            - boolean
            - ContentText: Text content.
              - …
            - ContentImage: Image content.
              - …
            - ContentAudio: Audio content.
              - …
            - ContentVideo: Video content.
              - …
            - ContentDocument: Document content (e.g. a PDF).
              - …
            - union[]
              - …
          - `span_id` string, nullable
          - `timestamp` string, required
          - `truncated` unknown[], nullable
            - unknown
          - `type` 'function', required
          - `uuid` string, nullable
          - `view` ToolCallContent: Content to include in tool call view.
            - `content` string, required
            - `format` 'text' | 'markdown', required
            - `title` string, nullable
          - `working_start` number, required
          - `working_time` number, nullable
        - AnchorEvent: Marks a rollback-able point in a replayable trajectory. Emitted by an orchestrator immediately after a step it can later roll back to (a staged message, a model generate, etc.). Carries no content; ``anchor_id`` is the addressable identifier that ``TimelineSpan.branched_from`` references. Hidden from the viewer transcript by default.
          - `anchor_id` string, required
          - `event` 'anchor', required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `source` string, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - ApprovalEvent: Tool approval.
          - `approver` string, required
          - `call` ToolCall, required
            - `arguments` object, required
            - `function` string, required
            - `id` string, required
            - `parse_error` string, nullable
            - `type` 'function' | 'custom', required
            - `view` ToolCallContent: Content to include in tool call view.
              - …
          - `decision` 'approve' | 'modify' | 'reject' | 'escalate' | 'terminate', required
          - `event` 'approval', required
          - `explanation` string, nullable
          - `message` string, required
          - `metadata` object, nullable
          - `modified` ToolCall
            - `arguments` object, required
            - `function` string, required
            - `id` string, required
            - `parse_error` string, nullable
            - `type` 'function' | 'custom', required
            - `view` ToolCallContent: Content to include in tool call view.
              - …
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `view` ToolCallView: Custom view of a tool call. Both `context` and `call` are optional. If `call` is not specified then the view will default to a syntax highlighted Python function call.
            - `call` ToolCallContent: Content to include in tool call view.
              - …
            - `context` ToolCallContent: Content to include in tool call view.
              - …
          - `working_start` number, required
        - ReviewEvent: Tool result review.
          - `call` ToolCall, required
            - `arguments` object, required
            - `function` string, required
            - `id` string, required
            - `parse_error` string, nullable
            - `type` 'function' | 'custom', required
            - `view` ToolCallContent: Content to include in tool call view.
              - …
          - `decision` 'continue' | 'terminate' | 'escalate', required
          - `event` 'review', required
          - `explanation` string, nullable
          - `message` string, required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `reviewer` string, required
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - SentinelEvent: Report from a sentinel monitor or protocol about one step. A report's metadata is recorded in the event's `metadata` field. Experimental: not yet a stable API; may change without notice.
          - `action` 'proceed' | 'modify' | 'reject' | 'terminate' | 'escalate', nullable
          - `audit` boolean, required
          - `conversation` string, required
          - `error` string, nullable
          - `event` 'sentinel', required
          - `explanation` string, nullable
          - `factory` string, required
          - `function` string, nullable
          - `kind` 'observation' | 'decision', required
          - `message` string, nullable
          - `metadata` object, nullable
          - `modified` ToolCall
            - `arguments` object, required
            - `function` string, required
            - `id` string, required
            - `parse_error` string, nullable
            - `type` 'function' | 'custom', required
            - `view` ToolCallContent: Content to include in tool call view.
              - …
          - `path` string, required
          - `pending` boolean, nullable
          - `references` Reference[], required
            - `cite` string, nullable
            - `id` string, required
            - `type` 'message' | 'event', required
          - `span_id` string, nullable
          - `stage` 'model_input' | 'model_output' | 'tool_call' | 'tool_result', required
          - `status` 'reported' | 'cancelled' | 'bypassed' | 'superseded' | 'error', required
          - `step_id` string, required
          - `suspicion` union
            - number
            - object
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - CommunicationEvent: Communication between agents: a record sent or delivered, a read of what others communicated, or a wait for others. Experimental: not yet a stable API.
          - `agent` string, required
          - `agent_span_id` string, nullable
          - `channel` string, nullable
          - `data` object, required
          - `delivery` CommunicationDelivery: For `delivered`, whom a record reached and in what order.
            - `recipients` string[], required
            - `seq` integer, nullable
          - `event` 'communication', required
          - `kind` 'sent' | 'delivered' | 'read' | 'waited', required
          - `loop_span_id` string, nullable
          - `metadata` object, nullable
          - `nested` boolean, required
          - `pending` boolean, nullable
          - `record_ids` string[], required
          - `send` CommunicationSend: For `sent`, the record an agent sent, and why it was refused if it was.
            - `record` CommunicationRecord, required: For `sent`, what one agent communicated: a message, or a change to state other agents read.
              - …
            - `rejected` CommunicationRejection: For `sent`, why a record that was sent was not delivered.
              - …
          - `source` string, required
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `wait` CommunicationWait: For `waited`, an agent waiting for other agents.
            - `channels` string[], required
            - `duration` number, required
            - `id` string, nullable
            - `outcome` string, required
          - `working_start` number, required
        - SwarmEvent: Evidence from an agent swarm (inspect_swarm) other than communication: a member's result, the final answer's selection, the swarm's ledger or its metrics. The swarm defines `kind` and the format of `data`. Experimental: not yet a stable API.
          - `data` object, required
          - `event` 'swarm', required
          - `kind` string, required
          - `member` string, nullable
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - BranchEvent: Marks where a branched trajectory's unique content begins. Emitted at the point where a branch transitions from replaying its parent's prefix to live execution. Events before this in the trajectory's span are replay-phase re-execution; events after are the branch's genuine new content.
          - `event` 'branch', required
          - `from_anchor` string, required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `span_id` string, nullable
          - `timestamp` string, required
          - `uuid` string, nullable
          - `working_start` number, required
        - CheckpointEvent: A successful checkpoint commit. Emitted by the checkpointer immediately after the per-checkpoint file JSON is written — see working.md §8a. Carries the full checkpoint payload flattened into top-level fields (via multiple inheritance from :class:`Checkpoint`), so a consumer of ``transcript().events`` (or the ``.eval`` log) reads ``event.checkpoint_id`` / ``event.trigger`` / ``event.host`` etc. directly — same data as someone reading ``<sample>/ckpt-NNNNN.json`` from disk.
          - `checkpoint_id` integer, required
          - `created_at` string, date-time, required
          - `duration_ms` integer, required
          - `event` 'checkpoint', required
          - `host` SnapshotDetails, required: Per-backup stats captured in the checkpoint file. One per repo (host repo + one per active sandbox repo). Values come from restic's backup summary — see :class:`ResticBackupSummary`.
            - `additional_files` integer, nullable
            - `duration_ms` integer, required
            - `files` string[], nullable
            - `size_bytes` integer, required
            - `snapshot_id` string, required
          - `metadata` object, nullable
          - `pending` boolean, nullable
          - `sandboxes` object, required
          - `size_bytes` integer, required
          - `span_id` string, nullable
          - `timestamp` string, required
          - `trigger` 'time' | 'turn' | 'manual' | 'token' | 'cost' | 'budget' | 'agent_complete', required
          - `trigger_metadata` object, nullable
          - `turn` integer, required
          - `uuid` string, nullable
          - `working_start` number, required
        - CompactionEvent: Compaction of conversation history.
          - `event` 'compaction', required
- … truncated; see the full OpenAPI document linked below

## Changes

- **2026-11-03** `b799e9a81653`: 3 info
  - added `#/components/schemas/CommunicationEvent, #/components/schemas/SwarmEvent` to the `input/anyOf[subschema #1: Transcript]/events/items/` response property `anyOf` list for the response status `200`
  - added `#/components/schemas/CommunicationEvent, #/components/schemas/SwarmEvent, subschema #6, subschema #34` to the `input` response property `anyOf` list for the response status `200`
  - removed `subschema #6, subschema #32` from the `input` response property `anyOf` list for the response status `200`
- **2026-10-27** `f854dcec6157`: 3 info
  - added the optional property `input/anyOf[subschema #12: ModelEvent]/origin` to the response with the `200` status
  - added the optional property `input/anyOf[subschema #1: Transcript]/events/items/anyOf[subschema #6: ModelEvent]/origin` to the response with the `200` status
  - added the optional property `input/anyOf[subschema #32]/items/anyOf[subschema #6: ModelEvent]/origin` to the response with the `200` status
- …earlier changes not shown

[Operation history](https://skmtc.dev/meridianlabs-ai/apis/inspect-scout-viewer-api/changes/schema/scanner-input/get.md)

---

[API](https://skmtc.dev/meridianlabs-ai/apis/inspect-scout-viewer-api.md) - [All operations](https://skmtc.dev/meridianlabs-ai/apis/inspect-scout-viewer-api/llms.txt) - [OpenAPI document](https://skmtc.dev/meridianlabs-ai/apis/inspect-scout-viewer-api/revisions/b799e9a81653?raw)