用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/neilsonnn/image-blaster --skill image-blast-sfx命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | image-blast-sfx |
| description | Generate and post-process a clean sound effect, object impact, or ambient loop. |
Generate an SFX for project $0.
$0 is missing, ask for the world slug.ls -a before reading generated state.worlds/$0/output/sfx/, use --loop --count 2 --kind world-ambience --prefix ambient-loop --duration-seconds 10.worlds/$0/output/<object-id>/sfx/, use --count 4 --kind object-impact --prefix impact-<object-id> --duration-seconds 1. Do not use --loop.worlds/$0/output/sfx/ unless an object is clearly specified, use --kind arbitrary.impact one-shot, short-decay, <object material description from object.json> hitting a hard surface.ambient environment, loop of <description of only ambient qualities of scene from image.json>.ffprobe/ffmpeg, trims leading/trailing silence or low noise, normalizes loudness, and stores audio_analysis in the hidden request JSON. Loop output is left as raw provider audio so the seamless boundary is preserved.node .claude/scripts/project/project-state.mjs --world "$0"
Run:
node .claude/scripts/sfx/fal-elevenlabs-sfx.mjs \
--prompt "<sound prompt>" \
--output-dir "<target output dir>" \
--prefix "<safe prefix>" \
--count "<1-4>" \
--kind "<world-ambience|object-impact|arbitrary>" \
--duration-seconds "<optional 0.5-22>" \
--loop "<for world ambience or explicit loop requests>" \
--postprocess true
Add --loop only for world ambience or explicit looping requests. Avoid music or voices unless explicitly requested.
If request metadata records provider URLs but local audio files are missing, fill them from the matching hidden request JSON:
node .claude/scripts/project/ensure-local-assets.mjs --from "<request-json-path>"
Final response: report generated audio files, loop status, request metadata, prompt used, and the trimming/quality notes from audio_analysis.