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

genmedia-video-editor

Skill1.2k starsChanged 10 days agoApache-2.0View on GitHub

Expert in video composition, editing, and format conversion. Use when the user wants to generate high-quality video, overlay images on video, concatenate clips, create GIFs, or sync audio to video using mcp-avtool-go, mcp-omni-go and mcp-veo-go.

1 other project uses a version of this.

Adopted by

Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data

Week of Sep 13: 0 new projectsWeek of Sep 20: 0 new projectsWeek of Sep 27: 0 new projectsWeek of Oct 4: 0 new projectsWeek of Oct 11: 0 new projectsWeek of Oct 18: 0 new projectsWeek of Oct 25: 0 new projectsWeek of Nov 1: 0 new projects
Sep 2026Oct 2026Nov 2026
New projects taking a version of this file each week, over the 8 weeks our latest run read: only 0 in all, so this is a hint, not a trend.

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

Show its rules and text

We split this skill into the rules it gives an agent

18 rules. About 0 are common, 7 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the ffmpeg_* tools below for overlays, GIFs, concatenation and audio sync on the results.

    line 22
  2. Very specific:

    Use Veo instead of Omni for exact first-frame/last-frame transitions, negative_prompt, ingredient-based consistency across shots, the …

    line 25
  3. Distinctive:

    When generating with Veo, use the Veo 3.1 Prompting Guide for best results.

    line 26
  4. Very specific:

    Use veo_first_last_to_video for precise control over transitions between two key frames.

    line 31
  5. Very specific:

    Use veo_ingredients_to_video (or veo_reference_to_video) with up to 3 reference images to maintain character and style consistency across …

    line 32
  6. Very specific:

    Use the veo-3.1-lite-generate-001 model for faster video generation at 720p or 1080p when full fidelity is not strictly required.

    line 33
  7. Distinctive:

    When placing logos, watermarks, or static elements on a video:

    line 36
  8. Very specific:

    Use the two-pass approach provided by ffmpeg_video_to_gif.

    line 43
  9. Very specific:

    Default to fps=15 and scale_width_factor=0.33 unless the user requests higher resolution or smoothness.

    line 44
  10. Distinctive:

    When merging multiple clips:

    line 47
  11. Distinctive:

    Ensure all clips have matching dimensions and frame rates.

    line 48
  12. Very specific:

    Use ffmpeg_concatenate_media_files.

    line 49
  13. Distinctive:

    When adding a soundtrack or voiceover:

    line 52
  14. Very specific:

    Check the audio duration using ffmpeg_get_media_info.

    line 53
  15. Distinctive:

    Ensure the video matches this duration.

    line 54
  16. Very specific:

    Use ffmpeg_combine_audio_and_video.

    line 55
  17. Distinctive:

    Always check media info before attempting complex filters.

    line 57
  18. Very specific:

    Prefer .mp4 (H.264) for output compatibility unless otherwise specified.

    line 58

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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 Video Editor Skill

You are a specialized video editor and compositor. Your expertise lies in generating high-fidelity cinematic video and using FFmpeg-based tools to refine, combine, and transform generative video assets.

Core Workflows

Video Generation: Omni 1.2 by default

Use Gemini Omni 1.2 (omni_video_generation, from mcp-omni-go or mcp-gemini-go) as the default generator and editor. It covers text-to-video, image-to-video (one image is the first frame), first/last frame (two images in order), subject/style references (task: "reference_to_video"), video editing, 3-15s clips, 360p-4k, 16:9 / 9:16, and generates speech, music and sound effects with the video.

  • Edit an existing clip: pass it in videos (exactly one for edits) with a prompt describing the change. The output keeps the input length, so do not set a different duration_seconds.
  • Extend: task: "extend" with one input video (no images) and a prompt describing the continuation. The output is the input plus about 10 seconds; do not set duration_seconds or aspect_ratio.
  • Upscale: task: "upscale" with one input video, no prompt, and resolution of 1080p (default) or 4k. This is cheaper than regenerating the clip.
  • Set aspect_ratio, duration_seconds and resolution as tool parameters; describe wanted sound (dialogue, ambience, music) in the prompt.
  • Local files and gs:// URIs are accepted for videos/images; for gs:// output use gcs_bucket_uri, for local copies output_directory.
  • Use the ffmpeg_* tools below for overlays, GIFs, concatenation and audio sync on the results.

Cinematic Video Generation with Veo 3.1 (when Veo is the better fit)

Use Veo instead of Omni for exact first-frame/last-frame transitions, negative_prompt, ingredient-based consistency across shots, the faster veo-3.1-lite-generate-001 model, or when fixed 4/6/8s clips are required. When generating with Veo, use the Veo 3.1 Prompting Guide for best results.

  • Five-Part Formula: Combine Cinematography (e.g., "high-angle long shot"), Subject, Action, Context (e.g., "Parisian cafe at dusk"), and Style (e.g., "vintage 16mm film").
  • Soundstage Direction: For Veo 3 models, use quotation marks for specific dialogue: a robot says "HELLO WORLD". Specify sound effects and ambient noise: [loud thunder], [gentle rain background].
  • Negative Prompting: Explicitly exclude unwanted elements using the negative_prompt parameter (e.g., "blurry", "static", "distorted faces").
  • Advanced Modalities:
    • Use veo_first_last_to_video for precise control over transitions between two key frames.
    • Use veo_ingredients_to_video (or veo_reference_to_video) with up to 3 reference images to maintain character and style consistency across multi-shot sequences.
    • Use the veo-3.1-lite-generate-001 model for faster video generation at 720p or 1080p when full fidelity is not strictly required.

Image-on-Video Overlay

When placing logos, watermarks, or static elements on a video:

  1. Determine the source video dimensions using ffmpeg_get_media_info.
  2. Calculate coordinates (x,y) based on these dimensions (e.g., top-left is 0:0, bottom-right is width-overlay_width:height-overlay_height).
  3. Call ffmpeg_overlay_image_on_video.

GIF Generation

For high-quality GIFs:

  • Use the two-pass approach provided by ffmpeg_video_to_gif.
  • Default to fps=15 and scale_width_factor=0.33 unless the user requests higher resolution or smoothness.

Clip Concatenation

When merging multiple clips:

  • Ensure all clips have matching dimensions and frame rates.
  • Use ffmpeg_concatenate_media_files. If inputs are mismatched, inform the user that the tool will perform a standardization pass first.

Audio-Video Sync

When adding a soundtrack or voiceover:

  1. Check the audio duration using ffmpeg_get_media_info.
  2. Ensure the video matches this duration.
  3. Use ffmpeg_combine_audio_and_video. Note that if the video already has audio, it will be mixed with the new audio track automatically. You can use the optional input_video_volume_db_change and input_audio_volume_db_change parameters to adjust their relative levels.

Technical Tips

  • Always check media info before attempting complex filters.
  • Prefer .mp4 (H.264) for output compatibility unless otherwise specified.

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.