comfyui-audio-creator（comfyui 音频创作者）

Provided by TippyEntertainment

https://github.com/tippyentertainment/skills.git

This skill is designed for use on the Tasking.tech agent platform (https://tasking.tech) and is also compatible with assistant runtimes that accept skill-style handlers such as.claude,.openai, and.mistral. Use this skill for both Claude code and Tasking.tech agent source.

comfyui-audio-creator

Summary

Generate background music and ambience that matches one or more video scenes produced by comfyui-video-generator. This skill does *not* handle spoken dialogue or narration; use comfyui-voice-generator for any voice work.

When to Use

• After video clips exist and the user wants music or ambient soundscapes for them.
• The user asks for background tracks for a trailer, promo, or game scene.
• You need loopable ambience (e.g., city at night, fantasy forest, sci‑fi engine room).

Inputs to Collect

Ask the user (or previous skill) for:

• Audio type

- music or ambience (no voice).

• Target scene(s)

- Reference to clips from comfyui-video-generator (IDs, durations, moods).

• Style / genre / mood

- Music: orchestral, synthwave, lofi, rock, cinematic, etc. - Ambience: forest, city, sci‑fi engine, horror, cozy home, etc.

• Duration

- Typically match the clip length or a loopable segment (e.g., 15–60 seconds).

• Loopability

- Whether the audio should seamlessly loop.

• Tempo / intensity (for music, optional)

- Slow/medium/fast; low/medium/high intensity.

• Format & sample rate

- Default: 16‑bit WAV, 44.1 kHz or 48 kHz, stereo.

Expected Behavior

1. Map the request into structured audio parameters:

- audioType, genre, mood, durationSeconds, loopable, targetScenes, tempo, intensity.

1. Select the appropriate ComfyUI or connected audio workflow:

- Music generator or ambience generator (no TTS here).

1. Generate audio in one or more segments:

- Ensure total duration covers the referenced scenes. - For loops, ensure clean loop points.

1. Collect:

- audioUrl / audioPath - durationSeconds - format, sampleRate, channels - Tags/metadata (genre, mood, bpm if known).

1. Return a concise summary plus technical details.

Output Format (to the caller)

Return a structure containing:

• description: e.g. “Loopable synthwave track for the opening scene.”
• audioUrl or audioPath
• durationSeconds
• format
• sampleRate
• channels
• loopable: boolean
• tags: list of strings (genre, mood, tempo)

Orchestration Notes

Typical pipeline:

1. comfyui-video-generator → create scenes.
2. comfyui-audio-creator → create background music/ambience for those scenes.
3. comfyui-voice-generator → add narration or dialogue.
4. comfyui-soundfx-creator → layer in moment‑specific sound effects.

This skill focuses on non‑voice background audio so other skills can handle speech and SFX cleanly.