Byted Airesearch Videoeval

Use this skill to submit and query long-running material evaluation tasks.

When to use

Use this skill when the user wants to:

• - evaluate a video material or creative asset
• submit a material evaluation task and get back a task identifier
• check the status of an existing evaluation task later
• fetch the final detail/result of a previously created evaluation task

Do not use this skill for generic video upload requests.

Current workflow

1. 1. Validate the full input batch before any upload starts.
2. Upload the local video files and capture the returned attachment_id values.
3. Create a task with the uploaded attachment IDs.
4. Return success immediately after the task is created.
5. Ask the user to query task list or task detail later if they want progress or results.

This workflow is intentionally non-blocking. Do not poll automatically after task creation.

Mandatory behavior

• - Do not expose the upload API as a standalone user-facing capability.
• Do not trigger this skill for generic requests such as “upload this video”, “store this file”, or “send this video”.
• Only call the upload API when the user explicitly intends to create a new material evaluation task.
• For new task creation, prefer scripts/submit_evaluation_task.py so validation, upload, and task creation stay in one controlled flow.
• Treat scripts/upload_video.py as an internal helper used by the orchestration flow, not as the primary user entrypoint. The script itself rejects direct use unless it is called with the internal orchestration marker.

For multi-file submissions, use the orchestration entrypoint so the whole batch is validated before the first upload starts.

Authentication

The current APIs use API key authentication.

All API requests sent by this skill must include the header:

• - INLINECODE3
• INLINECODE4

Preferred input methods:

• - INLINECODE5
• INLINECODE6

If no API key is available, ask the user to create or view one at:

• - INLINECODE7

Then ask the user to provide the API key before calling the API.

If the API key is missing, the scripts must fail immediately with a clear error that points the user to the API key page above.

Known API coverage

Upload attachment

• - Endpoint: INLINECODE8
• Request: INLINECODE9
• File field: INLINECODE10
• Upload constraints:

• - Output mapping:

The upload_video.py script is an internal helper for the create-task workflow. It is not the primary user-facing entrypoint.

Create task

• - Endpoint: INLINECODE16
• Fixed request fields:

• - Optional request fields currently exposed:

• - Create constraints:

The create step should send the uploaded attachment_id as a one-element attachment_ids array unless multiple attachment IDs are explicitly provided.

List and detail

Task detail is wired and can query an existing task directly:

• - Endpoint: INLINECODE30
• Current auth: API key bearer token
• Output mapping:

Task list is wired and can query existing tasks directly:

• - Endpoint: INLINECODE39
• Current auth: API key bearer token
• Current fixed query params:

• - Output mapping:

Commands

CODEBLOCK0

Response handling

All scripts emit JSON to stdout with the same top-level envelope:

• - INLINECODE49
• INLINECODE50
• INLINECODE51
• INLINECODE52
• INLINECODE53

Important normalized fields:

• - submit task: data.task_id, data.task_status, INLINECODE56
• list: INLINECODE57
• detail: data.detail, INLINECODE59

Final answer rules

• - Use the structured JSON returned by the detail endpoint as the internal source of truth.
• Present the final answer in human-readable natural language.
• When the task is finished, prefer a concise report-style answer rather than a raw data dump.
• Do not dump raw JSON to the user.
• Do not expose internal field names such as video_eval, video_user_report, distribution, field_desc, or similar implementation-oriented keys.
• Apply the same rule to task list responses: use the list result as internal source data, but present the outcome as a natural-language summary rather than raw fields.
• For task list answers, it is acceptable to include the task ID, task name, status, created time, and updated time in human-readable prose, because those fields help the user choose a task for follow-up detail queries.
• When multiple videos are present, summarize them separately.
• For finished task detail results, prefer a readable report flow such as: task conclusion first, then one short section per video, then overall recommendations if the source data supports them.
• If the task is not finished yet, do not fabricate a report. Clearly state the current status and ask the user to check again later.
• Only provide raw structured data if the user explicitly asks for the raw result.

Practical guidance

• - For new submissions, use the orchestration flow rather than exposing upload as a standalone step to the user.
• Validate the local file before upload. Reject non-MP4 files, files with non-video/mp4 MIME types, or files larger than 50MB with a direct and actionable error message.
• Validate the task creation input before calling the API. Reject any request that contains more than 50 attachment IDs with a direct and actionable error message.
• For a multi-video submit flow, validate the full batch size before any upload starts. If the batch contains more than 50 files, fail immediately and do not upload anything.
• After create succeeds, tell the user the task has been submitted successfully and can be checked later.
• Use task list when the user wants to browse or find historical tasks.
• Use task detail when the user already knows the task ID and wants the final result.