商品主图视频发布门禁

Use this skill after product-video-asset-audit to turn audit results into an actionable review queue for a human publisher.

Prerequisites

• - Python 3.9 or newer.
• A JSON audit produced from redacted facts.
• A platform requirement profile supplied by the owner or platform documentation.

Input and output

Input is the audit JSON plus a profile with duration, dimensions, audio, claims, rights, and disclosure requirements.

Output is productvideopublishgate.json. It records MANUALAPPROVALREQUIRED or HOLDREPAIRREVIEW, lists field-level tasks, and fixes manualpublishallowed=false and externalpublish_executed=false.

Procedure

1. 1. Run the asset audit and keep the output unchanged as the evidence record.
2. Run python scripts/buildproductvideopublishgate.py --audit productvideoaudit.json --output productvideopublish_gate.json.
3. Give the task list to the owner or publisher; make any repair in the owner\'s normal editing tool and rerun both skills.

Error recovery

Missing audit fields, malformed JSON, or an audit with no checks exits non-zero. The gate never converts unknown evidence into approval.

Permission and privacy boundaries

This gate does not upload, publish, approve, edit, remove watermarks, alter claims, or certify compliance. It does not access seller, customer, payment, order, or advertising data. A human must approve any irreversible external action.

Acceptance

The example audit is held because the audit contains a review item. A clean audit can produce MANUALAPPROVALREQUIRED with an empty repair list, because final publication remains a human action.