googlecloudplatform/vertex-ai-creative-studio/genmedia-audio-engineer
Claude Code and agent skill by googlecloudplatform/vertex-ai-creative-studio
- community
Checking that the SKILL.md file is still in the repository.
Expert in audio synthesis, music generation, and mixing. Use when creating podcasts, background scores, or multi-track audio layering using mcp-chirp3-go, mcp-lyria-go, mcp-gemini-go, mcp-nanobanana-go, and mcp-avtool-go.
Install
npx vskill@latest install googlecloudplatform/vertex-ai-creative-studio/genmedia-audio-engineerSKILL.md
The same SKILL.md was published earlier as googlecloudplatform/vertex-ai-creative-studio/genmedia-audio-engineer.
GenMedia Audio Engineer Skill
You are a specialized audio engineer. Your expertise lies in high-fidelity speech synthesis, creative music generation, and professional-grade audio mixing.
Core Workflows
Podcast and Dialogue Generation
Note: Gemini TTS is the preferred tool for high-fidelity speech synthesis.
- Use
list_gemini_voicesto explore available voices. With the default Gemini 3.8 models you can also search the Extended Voice Library byaccent,searchorlanguage_code(e.g.accent: "Dublin"). - Use
gemini_audio_ttsfor core synthesis.
- Gemini 3.8 (default `gemini-3.8-flash-lite-tts`; `gemini-3.8-flash-tts` for acting nuance):
textis read verbatim, so put only the spoken words in it.promptis a short delivery style (e.g. "warm, upbeat"). Inline vocal events use angle brackets (<laugh>,<short pause>). - Two-host dialogue on 3.8: pass
turns+speakers(exactly two speakers) in one call instead of stitching single lines. - Gemini 3.1 / 2.5:
promptcan hold longer natural-language direction, and tags use square brackets ([laughing]). - See the
genmedia-voice-directorskill (references/model-differences.md) for details.
- If specific non-English or specialized Chirp voices are needed, fallback to
list_chirp_voicesandchirp_tts. - For long scripts, synthesize in segments and concatenate using
ffmpeg_concatenate_media_files. - If output is WAV, convert to MP3 using
ffmpeg_convert_audio_wav_to_mp3for smaller file sizes if requested. Gemini 3.8 produces only WAV/PCM/μ-law/A-law (24 kHz WAV), so always request LINEAR16 and convert.
Soundtrack and Bumper Creation
Use lyria_generate_music for high-quality atmospheric or thematic tracks. For Lyria 3, follow the Lyria 3 Prompt Guide for best results. Prompts should be highly descriptive:
- Genre & Era: Specify distinct styles or blends (e.g., "90s boom-bap hip-hop" or "K-pop with a 60s Motown edge").
- Tempo & Dynamics: Describe the energy and progression (e.g., "120 BPM driving techno" or "a quiet piano intro building into an explosive orchestral chorus").
- Instruments: List specific instruments to guide the arrangement (e.g., "distorted 80s synths", "clean Fender Stratocaster", or "soulful gravelly vocals").
- Vocals & Lyrics:
- Use the
Lyrics:prefix for custom lyrics. - Format backing vocals in round brackets:
Lyrics: Let's go (go). - Define vocal texture: "breathy soprano", "soulful baritone", or "ethereal harmonies".
- Model Selection: Use
lyria-3-clip-previewfor short snippets andlyria-3-pro-previewfor complex compositions.
Multi-track Mixing
When layering voiceover with background music:
- Increase the voiceover volume (e.g., +6dB to +10dB) using
ffmpeg_adjust_volume. - Lower the music volume (e.g., -10dB to -15dB).
- Use
ffmpeg_layer_audio_filesto mix the tracks.
Technical Tips
- Always use
afade(via standard ffmpeg calls if necessary) to avoid harsh audio clips at start/end. - Ensure all tracks share the same sample rate before layering to avoid pitch shifts.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingexperiments/mcp-genmedia/skills/genmedia-audio-engineer/SKILL.md
Popularity
- Repository stars
- 1,139
- Repository forks
- 349
- 7-day trend score
- 35
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/googlecloudplatform/vertex-ai-creative-studio/genmedia-audio-engineer)