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.
git clone --depth 1 https://github.com/open-video-ai/open-videonpx agentmods add skills/open-video-ai/open-video/h3-videoWrote 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.
[](https://agentmods.dev/skills/open-video-ai/open-video/h3-video)<a href="https://agentmods.dev/skills/open-video-ai/open-video/h3-video"><img src="https://agentmods.dev/badge/skills/open-video-ai/open-video/h3-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.
<a href="https://agentmods.dev/skills/open-video-ai/open-video/h3-video"><img src="https://agentmods.dev/badge/skills/open-video-ai/open-video/h3-video.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00103 | $0.03380 |
| Opus 5 | $0.00051 | $0.01690 |
| Sonnet 5 | $0.00021 | $0.00676 |
| Haiku 4.5 | $0.00010 | $0.00338 |
Grade C, and why
h3-video scanned grade C with 2 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 12d ago.
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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
curl -fsSL https://open-video.ai/install | bash # Linux/macOS Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sf http://127.0.0.1:8188/system_stats || ( How it starts
The opening of the file, as written. The whole thing — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenVideo skill · v0.0.1
Brand: OpenVideo (always lead with this). MiniMax H3 is the model OpenVideo drives.
Job: use OpenVideo so an agent produces good product video, not a random diffusion click. Quality comes from prompt craft + correct mode + validated settings, not secret samplers.
0. Paths (env-first)
Recommended sibling layout (any parent dir name; no machine-absolute paths):
parent/
├── open-video/ # THIS product — CLI, backends, skills ← work here
└── lab/ # ComfyUI + weights (not git)
| What | Path / env |
|---|---|
| Product root (CLI, skills) | $OPEN_VIDEO_ROOT (this checkout) |
| Install / pull / run | From product root: ./scripts/install.sh, python -m open_video … |
| Lab engine + weights | $OPEN_VIDEO_LAB / $H3_LAB (sibling lab/ with ComfyUI + h3_models/) |
| Weights env | export OPEN_VIDEO_MODELS=$H3_LAB/h3_models |
| ComfyUI URL | export OPEN_VIDEO_COMFYUI=http://127.0.0.1:8188 (default) |
Resolve product root (never hardcode a machine path):
# Prefer env, else directory of this skill → repo root
export OPEN_VIDEO_ROOT="${OPEN_VIDEO_ROOT:-$(cd "$(dirname "$0")/../.." 2>/dev/null && pwd)}"
cd "$OPEN_VIDEO_ROOT"
Lab harness (same ComfyUI) when you keep a sibling runtime:
export OPEN_VIDEO_LAB="${OPEN_VIDEO_LAB:-$OPEN_VIDEO_ROOT/../lab}"
export H3_LAB="${H3_LAB:-$OPEN_VIDEO_LAB}"
export OPEN_VIDEO_MODELS="${OPEN_VIDEO_MODELS:-$H3_LAB/h3_models}"
Product = open-video/ checkout. lab/ is runtime only — never the product root.
1. Quality hierarchy (what actually moves the needle)
| Rank | Lever | Rule |
|---|---|---|
| 1 | 3-field prompt | Official structure; concrete visible/audible detail; camera type+amplitude+speed; dialogue in <d>[lang]…</d> |
| 2 | Mode | T2V / I2V / FL2VA chosen correctly from inputs |
| 3 | Resolution / steps | 1344×768, 20 steps, res_multistep + simple (defaults in harness) |
| 4 | Quant | INT8 on 5090-class; lower VRAM → recommend-quant (nf4/w4/int8+lowvram) |
| 5 | Duration | 5–10 s sweet spot (max 15 s / shot); multi-shot via cut times or open-video director skill |
| 6 | Review | Watch / extract frames; fix prompt; re-run. Dual vision gate only when shipping |
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.
- 12d ago First seen · 307 lines · 103 tokens per session scan C 8383e461a3b4
h3-video is a skill published in the GitHub repository open-video-ai/open-video (117 stars, last pushed 24d ago), licensed Apache-2.0. It adds 103 tokens to every session and 3,380 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
video-gen
Use when generate videos with AI models — Runway, Kling, Sora, Pika, Seedance 2.0, Grok Imagine, Veo. Text-to-video, image-to-video, video extension, multi-modal references. Use when generating video from text prompts, animating images, or creating AI video content.
videoagent-video-studio
Generate short AI videos from text or images — text-to-video, image-to-video, and reference-based generation — with zero API key setup. Use when the user wants to create a video clip, animate an image, or generate video from a description.
flash-reel
Generates a 30-second 9 16 cinematic reel from a user prompt (setting + vibe + rough storyline). Flash Reel — 30s Cinematic Reel Skill When to use User provides: a setting, a vibe/mood, and optionally a rough storyline. They want a 30 second 9:16 reel with hard cuts every 3-4 seconds, anchored by one or more…
static-ads
Run when the user wants to recreate a winning ad format with their own products and brand copy, or use the /static-ads command. Takes an uploaded ad format reference image, derives the layout structure and copy framework internally, generates on-brand copy variations, then renders static ads via GPT-image-2 using…
kling-3-prompt-director
Production-ready Kling 3.0 video prompt director using the canonical 9-field formula. Includes locked character/environment specs for the Crococopter and Swa & Danny universes. Trigger on any request pairing "Kling" with prompt-related verbs.
seedance-prompting-skills-for-cinematic-films
When to use Trigger this skill when the user requests photorealistic cinematic motion — not stylized animation, not motion design, not UGC. Specifically.