video_review
Plan, acknowledge, and finish host-chosen visual evidence scopes with bounded image receipts, enabling agents to track inspected frames and resume pending batches after failures.
Instructions
Manage the HOST-CHOSEN evidence scope and bounded image receipts. action=status returns the plan, acknowledged indices, pending batch and nextAction. plan: visual=all means all generated images for this task; selected requires indices; none permits text-only evidence. Text processing status, coverage and pagination remain available through transcript_read, but text reading is not tracked or enforced by this review. There is no fixed global completeness or ASR requirement. ack: after actually inspecting a returned batch, explicitly supply batchId and viewedIndices; [] releases the batch without claiming inspection. Unchosen images may be skipped. replay returns the same pending batch after transport failure. reviewId resumes this consumer ledger after server restart; omitting it creates a fresh consumer when not already active. finish checks only the selected plan; it returns finished=false with missing work instead of inventing completeness. The Agent must assess adequacy to the user, keep its own semantic notes, and distinguish actual source coverage from acknowledged images. No descriptions or summaries are generated here.
The HOST chooses evidence depth from the user request and actual subtitle/ASR quality and coverage: text alone, selected storyboards, all storyboards, or targeted higher-quality follow-up. Subtitles, ASR and images are independently optional; choose any combination. For text-only media use audio_start/audio_url_start (audio-bearing video files/pages are accepted) without a visual job. No mandatory image retrieval, all-images rule, fixed answer format or mandatory ASR. Use video_review action=plan to record visual=none|selected|all; revise as evidence changes. Missing subtitles do NOT block usable images, and a video-only stream does not prove missing source audio. Each image response has a pending batch. Return images AND timestamp metadata to the model, inspect the chosen images, then call video_review action=ack with batchId and explicit viewedIndices (empty [] releases uninspected images without credit). ONLY THEN fetch another batch. Never accumulate many video_read calls inside one code/exec result: return control to the model after each bounded batch. Transport errors leave the batch pending: video_review replay retries it; status/reviewId support resume. Acknowledgement is a caller assertion, not proof of comprehension. Keep any semantic working notes in the Agent workspace, not in MCP evidence. Before finishing, video_review finish checks only the HOST-CHOSEN visual scope and pending image receipts, not text reading completion. The host must still judge if that scope answers the user, and state relevant gaps. Honest incomplete-work disclaimers do not justify abandoning feasible work. session_stop cleanup cancels processing but does not complete reading; user_cancelled is only for an explicit user cancellation. Audio-only workflows remain independent.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"id": "<id>"
},
"name": "video_review"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | ||
| action | No | ||
| visual | No | ||
| batchId | No | ||
| indices | No | ||
| reviewId | No | ||
| viewedIndices | No |