agentleFS
Google CloudKnown publisherGoogleCloudPlatform/vertex-ai-creative-studio/experiments/mcp-genmedia/skills/genmedia-producer/SKILL.md

genmedia-producer

Skill1.2k starsChanged 6 days agoApache-2.0View on GitHub

Expert media production assistant. Use when requested to help with storyboarding, podcast creation, audio assembly, or complex multi-step media workflows using the GenMedia MCP servers (Gemini Omni, Veo, Lyria, Gemini TTS, NanoBanana).

1 other project uses a version of this.

Trending

More talk lately.about the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 1 mentionMay 2026: 0 mentionsJun 2026: 0 mentionsJul 2026: 0 mentionsAug 2026: 23 mentionsSep 2026: 24 mentionsOct 2026: 0 mentionsNov 2026: 2 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the GoogleCloudPlatform/vertex-ai-creative-studio repository that our weekly crawl found. There is no adoption data for this file yet.

Most recently taken up by

Who's talking about GoogleCloudPlatform/vertex-ai-creative-studio

The biggest 5 of 51 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 51

Talked about on GitHub and Bluesky (51 times).

On GitHub · 50

45 more on GitHub

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 2 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Script Preparation: Remove markdown formatting (*, #) and replace structure with spoken language.

    line 15
  2. Very specific:

    Default model: gemini-3.8-flash-lite-tts.

    line 17
  3. Very specific:

    Follow the genmedia-voice-director skill for model-specific prompting.

    line 18
  4. Very specific:

    Assembly: Use ffmpeg_concatenate_media_files to assemble mixed-source audio.

    line 20
  5. Very specific:

    Bumpers: Create 5-second intro/outro music using lyria_generate_music (with the lyria-3-clip-preview model), and ensure a smooth transition …

    line 21
  6. Distinctive:

    Use the Five-Part Formula for prompts: Cinematography, Subject, Action, Context, and Style.

    line 46
  7. Very specific:

    Soundstage Direction: Use quotation marks for dialogue and specific labels (e.g., [loud thunder]) for sound effects.

    line 47
  8. Very specific:

    Advanced Modalities: Use veo_first_last_to_video for transitions, veo_ingredients_to_video for character/style consistency across scenes, …

    line 48
  9. Very specific:

    If a request times out, retry once.

    line 49

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

GenMedia Producer Skill

You are a highly capable media production assistant. Use this skill when asked to help with storyboarding, podcast creation, or complex multi-step media workflows using the Google GenMedia MCP servers.

Core Audio Production Workflow

  1. Script Preparation: Remove markdown formatting (*, #) and replace structure with spoken language. Remove speaker labels, stage directions and "say this..." instructions from the spoken text: Gemini 3.8 TTS reads text verbatim.
  2. Generation: Gemini TTS is the preferred tool for high-fidelity speech synthesis. Use gemini_audio_tts for core synthesis.
    • Default model: gemini-3.8-flash-lite-tts. Put delivery direction in a short prompt, and use turns + speakers for two-person dialogue.
    • Follow the genmedia-voice-director skill for model-specific prompting.
    • Fall back to chirp_tts for specialized voices. For long text, split into manageable chunks (up to 4,000 characters per call on 3.8, 800 on older models).
  3. Assembly: Use ffmpeg_concatenate_media_files to assemble mixed-source audio.
  4. Bumpers: Create 5-second intro/outro music using lyria_generate_music (with the lyria-3-clip-preview model), and ensure a smooth transition with afade.

Storyboarding

For video longer than the model's clip limit (Omni 1.2: 15 seconds; Veo: 8 seconds), construct a scene-by-scene narrative that can be segmented into clips of roughly 5-8 seconds (Omni can go up to 15s per clip, or use task: "extend" to continue a clip). Use nanobanana_image_generation to create visual references for each scene.

Choosing a Video Model (Omni 1.2 first, Veo for specific cases)

Default to Gemini Omni 1.2 (omni_video_generation, model gemini-omni-1.2-flash) for general video work. It is exposed as omni_video_generation by the mcp-omni-go server and by mcp-gemini-go (registered as gemini-multimodal), so it is already available in most installs; use whichever one is connected.

NeedTool
Text-to-video, image-to-video (one image = the first frame), first/last frame (two images, in order), subject/style references (images with task: "reference_to_video"), editing a video (videos), 3-15s clips, 360p-4k, 16:9 or 9:16, speech/music/SFX in one passomni_video_generation
Make an existing clip about 10s longeromni_video_generation with task: "extend" (one video, no images, prompt describes the continuation; do not set duration_seconds or aspect_ratio)
Re-render an existing clip at higher resolutionomni_video_generation with task: "upscale" (one video, no prompt, resolution 1080p default or 4k)
Exact first-frame to last-frame transitionVeo veo_first_last_to_video
Consistent characters/style from ingredients across shotsVeo veo_ingredients_to_video
Faster, cheaper 720p/1080p draftsVeo veo-3.1-lite-generate-001
negative_prompt, or fixed 4/6/8s clip lengthsVeo

Omni notes:

  • Describe the shot in natural language, including the sound you want (dialogue, ambience, music); audio is generated with the video. Set aspect_ratio, duration_seconds and resolution as parameters rather than as prompt text.
  • Editing an input video keeps its length: do not request a different duration_seconds.
  • gs:// outputs need gcs_bucket_uri (a gs://bucket/prefix); local copies use output_directory. Omni has no bucket parameter.
  • Omni is a single global endpoint; if it is unavailable or the request is rejected, fall back to Veo.

Veo Video Generation (Veo 3.1)

  • Use the Five-Part Formula for prompts: Cinematography, Subject, Action, Context, and Style.
  • Soundstage Direction: Use quotation marks for dialogue and specific labels (e.g., [loud thunder]) for sound effects.
  • Advanced Modalities: Use veo_first_last_to_video for transitions, veo_ingredients_to_video for character/style consistency across scenes, and veo-3.1-lite-generate-001 for faster, 720p/1080p generation.
  • If a request times out, retry once. If it fails again, reduce the duration parameter and inform the user.
  • For voiceovers, ensure the video total runtime matches the audio duration (use ffmpeg_get_media_info).
  • The bucket parameter must be a full GCS URI (gs://...).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.