Generative-Media-Skills is a toolkit that lets AI agents generate, edit, and display images, videos, and audio through the muapi command-line interface. It is for users of Claude Code, Cursor, Gemini CLI, and OpenCode who need multimodal media-generation workflows. The catalogue entries are the skills that expose these media capabilities to coding agents.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add SamurAIGPT/Generative-Media-Skills --skill music-videogit clone --depth 1 https://github.com/SamurAIGPT/Generative-Media-SkillsWrote 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/samuraigpt/generative-media-skills/music-video)<a href="https://agentmods.dev/skills/samuraigpt/generative-media-skills/music-video"><img src="https://agentmods.dev/badge/skills/samuraigpt/generative-media-skills/music-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/samuraigpt/generative-media-skills/music-video"><img src="https://agentmods.dev/badge/skills/samuraigpt/generative-media-skills/music-video.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 45 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Data Exfiltration · line 60 Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
- medium Data Exfiltration · line 60 Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
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 | $0.00027 | $0.00700 |
| Opus 5 | $0.00014 | $0.00350 |
| Sonnet 5 | $0.00005 | $0.00140 |
| Haiku 4.5 | $0.00003 | $0.00070 |
Grade B, and why
muapi-music-video scanned grade B 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 13d 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.
Sends data to an external URLmediumData exfiltration
A POST to an outside endpoint may be telemetry or may be exfiltration; either way the mod talks to somewhere, and you should know where.
- For model IDs without a CLI alias yet, fall back to the raw endpoint via `curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'` and poll with Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- For model IDs without a CLI alias yet, fall back to the raw endpoint via `curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'` and poll with How it starts
The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Music Video
Build a short music video from a song theme — N keyframes, animate each, generate matching music.
Inputs
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
theme |
text | yes | — | Song / video theme (e.g. "lonely robot finds a friend, hopeful"). |
scenes |
int | no | 3 | Number of scenes (each becomes a 5s clip). |
music_style |
text | no | ambient cinematic, instrumental, slow tempo, warm | Suno-style tags for the soundtrack. |
visual_style |
text | no | cinematic, photoreal, soft volumetric light, 16:9 |
Steps
Build one the plan covering:
- Layer A (parallel) — N keyframes + 1 music track all at once.
- For each scene 1..N:
muapi image generatewith a beat-specific prompt +{{visual_style}}, model=nano-banana-pro (these feed video gen). - One
muapi audio create(kind=music) using{{music_style}}, duration = N × 5 + a 2s tail.
- For each scene 1..N:
- Layer B (parallel, depends on Layer A) — animate each keyframe.
- For each scene:
muapi video from-imagewithimage=$nX.url, model=veo3.1-image-to-video, duration=5, prompt=scene-specific motion direction.
- For each scene:
- Return:
- The scene keyframes (asset ids in order).
- The animation clips (asset ids in order).
- The music track asset id.
- A short summary describing the cut order.
Notes
- Keep character continuity by repeating the character description in every scene prompt verbatim.
- Don't auto-confirm any single video call > 50 cr — those need the user's nod (the loop will prompt automatically).
- If a scene's
muapi video from-imagefails after failover, fall back tomuapi video generate(text-to-video) for that scene only.
Trigger Keywords
music video, mv, video story, song visualization
Notes for the Executing Agent
- This recipe is LLM-orchestrated: read each phase, gather any missing inputs from the user, then call
muapiCLI commands. Usemuapi auth configurefirst ifMUAPI_API_KEYis unset. - For model IDs without a CLI alias yet, fall back to the raw endpoint via
curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'and poll withmuapi predict wait <request_id>. - Substitute
{{input_name}}placeholders with the user's actual inputs before issuing each call.
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.
- 13d ago First seen · 62 lines · 27 tokens per session scan B 95aa5c9a734b
muapi-music-video is a skill published in the GitHub repository SamurAIGPT/Generative-Media-Skills (4,263 stars, last pushed 3d ago), licensed MIT. It adds 27 tokens to every session and 700 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 2 findings (sends data to an external url, 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
mlops-automation
Automate an MLOps project with mise tasks, lefthook hooks, Docker images, GitHub Actions, and MLflow tracking on a SQL backend. Use when adding a task runner, git hooks, CI/CD, or experiment tracking to a working package.
mlops-validation
Add the validation layers that gate a merge — ty typing, Ruff linting, pytest coverage, structured logging, and the trivy, pip-audit, and gitleaks scans. Use when hardening code quality or wiring the mise run check task.
mlops-prototyping
Structure reproducible Jupyter notebooks with a fixed section layout, hoisted configuration, and leakage-free scikit-learn pipelines. Use when exploring a dataset, training a first model, or preparing a notebook for promotion.
mlops-collaboration
Prepare a project for public collaboration — license, code of conduct, docs, branch rulesets, templates, and git-cliff releases. Use when open-sourcing a repository, onboarding contributors, or cutting a tagged release.
mlops-observability
Make an ML system a glass box with reproducible runs, MLflow dataset lineage, drift monitoring, alerting, and SHAP explanations. Use when a deployed model needs traceability, monitoring, alerting, or explanation.
mlops-industrialization
Convert notebook prototypes into a distributable Python package with a src layout, a domain/io/application split, and validated OmegaConf plus Pydantic configuration. Use when moving code out of notebooks or designing entrypoints.