video-finishing
Normalize or deliberately reframe selected picture media for edit and delivery, with optional approved AI-upscale routing and new media identities for every finished derivative.
SKILL.mdskills/video-finishing/SKILL.md
GitHubVideo Finishing
Workflow
- Read
../../references/VIDEO_FINISHING.md, selections, media QC, media registry, and target timeline settings. - Decide whether a selected take needs no separate finish, conventional normalization, or an approved generative upscale.
- For conventional finishing, write
05_post/video_finish.jsonland usescripts/video_finish.py. - For an AI upscale, route through the declared production capability and matching model adapter. Preserve the original selected take.
- Register every derivative as a new media candidate.
- Re-run media QC after generative or materially destructive processing.
Done
Every picture file entering the final timeline either satisfies the timeline contract directly or has a traceable finished derivative.