Table-Video SOP skill
A Codex skill that runs AI video production in stages, from script and voiceover to first-cut review and final assembly.
- 289 ★ on the repository
- Last commit
Minimal access
Why this level
- The skill reads project files and writes progress files and drafts.
Install
- Works in
Manual install
Скопируйте папку skills/table-video в .agents/skills/table-video вашего видеопроекта
Codex install from the README: copy the folder whole so `SKILL.md` sits directly inside, and open Codex from that project.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill reads project files and writes progress files and drafts.
This is third-party code. Review the repository files before installing.
What it does
Table-Video SOP packages a video production process into a `table-video` skill: the agent reads material stage by stage, organizes work and records progress. The stages are script and voiceover, references and style, agreement on segments, shots and design, a first cut, human review with revisions, and final assembly. The repository holds rules, templates and prompts that the skill reads itself so you do not copy them by hand. Confirmed decisions can be reused, but key plans and the finished video are still approved by a person. The author warns that a finished result on the first try is not expected, so leave time for manual review.
Who it is for. Creators of short explainer and tutorial videos who make clips with AI and want to keep the process in repeatable stages.
Good fit when
- A video script and voiceover exist and a shot plan is needed.
- A first cut is reviewed and revised from notes.
- A series of videos shares one style and common rules.
- Progress by stage must be recorded so work can resume later.
Not a fit when
- You expect a finished clip in one request with no review.
- Codex is not available.
- The video is made without AI tools.
Example request
Use the table-video skill for this episode. Script: scripts/ep12.md; voiceover: audio/ep12.wav; references: none yet.Limitations
The README is in Chinese, with an English version. The skill targets Codex and installs by copying a folder. Do not expect a result on the first try: watch the whole video at normal speed and send back revisions. Templates and prompts depend on generation and editing services that differ in availability by country, so check that up front. The repository carries two licenses, one for materials and one for code. The last commit was on 6 November 2026. Work runs through your own accounts and keys.
How to disable. Delete the `.agents/skills/table-video` folder in the project and restart Codex. Remove progress files and video drafts separately.
FAQ
Which agent is needed?
Codex: the skill goes into `.agents/skills/table-video`.
What input is needed?
A script, a voiceover if there is one and references if any.
Can I work without the skill?
Yes, the SOP and prompts in the repo can be followed by hand.
Who approves the result?
A person: key plans and the finished video are reviewed by hand.
Is there an English README?
Yes, `README.en.md`.
Discussion
No questions or reviews yet.
Related
Skillsyt-dlp is a media-download CLI that lists available formats and simulates a download before saving media files to disk.
Whisper is a CLI for audio transcription, with model sizes for different hardware and output in the original spoken language.
Skills and CLI from HeyGen: the agent writes HTML compositions and renders them into MP4 videos, captions, motion graphics and decks
Agentic video studio for Claude Code, Cursor and Codex: script, voiceover, images, stock footage, music, subtitles and render
whisper.cpp is an audio CLI without a Python environment, compiled for a selected processor or hardware accelerator.
GPT-Image2 Style Library
gpt-image-2-style-library
A skill picks a matching template and style from a library of GPT-Image2 cases and assembles it into a finished prompt
README in short
The SOP splits AI video production into stages with confirmations; the skill reads rules and prompts itself and keeps a progress record.