A directing workflow for generating computer-created game video clips from text, starting images, or transitions between images. It prepares one validated set of shot instructions for each clip, but does not generate the video itself.

Use it to prepare opening scenes, cutscenes, ultimate attacks, or promotional clips for a game's video-generation pipeline.

It turns a game-video idea into a clear, model-specific task with a purpose, visual direction, duration, format, references, and acceptance criteria. This reduces ambiguity before video generation begins.

cg_video — Skill for Claude Code and Codex by OpenDCAI/GameFactory-3A

Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo

git clone --depth 1 https://github.com/OpenDCAI/GameFactory-3A

agentmods

npx agentmods add skills/opendcai/gamefactory-3a/cg_video

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/opendcai/gamefactory-3a/cg_video/github.svg)](https://agentmods.dev/skills/opendcai/gamefactory-3a/cg_video)

Your own site

<a href="https://agentmods.dev/skills/opendcai/gamefactory-3a/cg_video"><img src="https://agentmods.dev/badge/skills/opendcai/gamefactory-3a/cg_video/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/opendcai/gamefactory-3a/cg_video"><img src="https://agentmods.dev/badge/skills/opendcai/gamefactory-3a/cg_video.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 0 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 3,533 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00000	$0.03533
Opus 5.5	$0.00000	$0.01413
Sonnet 5.5	$0.00000	$0.00707
Haiku 4.5	$0.00000	$0.00353

Measured 1mo ago against content hash 2c662c4202dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

cg_video scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.

The scan reads SKILL.md. This mod also ships 1 executable file (game-cg-director/scripts/validate_output.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agent_skills/asset_qa/cg_video/SKILL.md · 370 lines

How it starts

The opening of the file, as written. The whole thing — 370 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CG-video generation and QA Skill

Use this Skill for text-to-video, first-frame-to-video, first/last-frame transitions, and reference-image-conditioned CG-video generation. Generate video only after the game plan defines its narrative purpose, visual style, shot, duration, and acceptance criteria.

Chain, boundaries, and artifacts

Game plan → game-cg-director → cg_tasks.jsonl
→ GenCGVideoOperator → VideoGenerationInput → video backend
→ MP4 bytes → video.mp4 + meta.json

Layer	Location	Responsibility
Model	`<REPO_PATH>/models/gen_cg_video/`	Model-native inference, cloud transport, and lifecycle
Operator	`<REPO_PATH>/operators/gen_cg_video/`	Task fields, local image loading, artifact paths, and metadata
Pipeline	`<REPO_PATH>/pipeline/assets_gen/gen_cg_video/`	Backend selection, CLI, JSONL batches, and summaries
Harness	`<REPO_PATH>/test/harness/`	CPU-only, network-free chain validation
Director sub-Skill	`<REPO_PATH>/agent_skills/asset_qa/cg_video/game-cg-director/`	Model-specific storyboard prompts and validated task rows

Paths in this Skill are written from the repository root. Resolve every <REPO_PATH>/... from there, and run every bash and python command below with <REPO_PATH>/ as the working directory.

Standard artifact layout:

test_data/outputs/<game_id>/<run_id>/assets/cg_video/<task_id>/
├── video.mp4
└── meta.json

A game that plays a clip through the Browser Serving gateway loads it by task identity at play time, and that request serves a stored artifact rather than creating one. Generate every clip named in the game plan during production, and run the play-time gateway with A3GAME_BROWSER_CG_VIDEO_PREBUILT_ONLY=1, so a clip that was never generated is reported as failed instead of being generated inside a page load. Where the served bytes have to stay fixed, pin the producing run_id: an unpinned lookup answers with the newest matching artifact.

Do not pass local image paths into a model directly. The task/JSONL uses local paths; the operator resolves them and converts images to PIL.Image.Image. Python callers with images already in memory may pass image objects, but never provide both an image object and its corresponding path field.

Read the full file on GitHub · 370 lines

Files

What ships with it

17 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

1mo ago Changed · +24 lines 2c662c4202dd
2mo ago First seen · 346 lines · 0 tokens per session scan A 479200de9e8e

Subscribe to this mod's changes

cg_video is a skill published in the GitHub repository OpenDCAI/GameFactory-3A (1,013 stars, last pushed 10d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,533 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

3d-generation Creative and media audio-generation blender image-generation motion-generation three-js

cg_video next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
cg_videoOpenDCAI/GameFactory-3A	Skill	1.0k	10d ago	Apache-2.0	A	0	Claude Code, Codex
agent-image-generationdanielgwilson/luxin	Skill	2	4d ago	MIT	A	163	Codex
luxindanielgwilson/luxin	Skill	2	4d ago	MIT	A	180	Claude Code
alpha-scene-genig-shadow-walker/BlenderXAlpha-3DGenSkill	Skill	31	3mo ago	MIT	A	284	Claude Code
image-generationdanielgwilson/luxin	Skill	2	4d ago	MIT	A	166	Codex
image-skilldanielgwilson/luxin	Skill	2	4d ago	MIT	A	165	Codex
creative-mediadanielgwilson/luxin	Skill	2	4d ago	MIT	A	141	Codex