agentleFS
bitwize-music-studiobitwize-music-studio/claude-ai-music-skills/skills/sound-effects/SKILL.md

sound-effects

Skill668 starsChanged 10 months agoCC0-1.0View on GitHub

Makes a sound effect, a one-shot or a loop for an album through Suno's MCP server - rain for an intro, a riser, a vinyl crackle bed, a drum loop at a track's tempo. Shows the cost and waits for a go-ahead. Use when the user asks for "a sound of...", "an ambience", "a loop", "a sting" or "a sound effect" rather than a song.

12 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use pose-library-design, video-toolkit and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • LinHao-citySkill

    pose-library-design

    Design reusable 2D character pose libraries, action cycles, and expression states for data-driven animation.

    Also in 4 projects that use this

  • LinHao-citySkill

    video-toolkit

    Create professional videos autonomously using claude-code-video-toolkit — AI voiceovers, image generation, music, talking heads, and Remotion rendering.

    Also in 4 projects that use this

  • GopyrSkill

    azure-speech-to-text

    Also in 3 projects that use this

  • GopyrSkill

    dashscope

    Also in 3 projects that use this

  • RobStrayerSkill

    voice-isolator

    Remove background noise and isolate vocals/speech from audio using ElevenLabs Voice Isolator (audio isolation) API. Use when cleaning up noisy recordings, removing music or background ambience from dialogue, isolating speech from field recordings, preparing audio for transcription, extracting vocals, or any "denoise / clean up / isolate voice" task.

    Also in 3 projects that use this

  • +7 more from LinHao-city

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

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

    Follow § Preflight in the reference doc: find sounds-create, audio-wait and account-usage, load their current definitions, and read …

    line 40
  2. Distinctive:

    Leave out Fitted when nothing is fitted.

    line 64
  3. Very specific:

    Call sounds-create with the keys of request and the idempotency_key.

    line 69
  4. Very specific:

    Call account-usage again. Credits spent = balance before − balance after.

    line 72
  5. Very specific:

    If the user wonders what one of them holds, Suno's listening agent says for free: audio-listen(id, question: "What is heard in this clip?

    line 87

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

Your Task

Input: $ARGUMENTS

Make a short sound for an album through Suno's MCP server: build the request, show what it costs, get a go-ahead, send it, file the two sounds, and let the user hear them.

You never choose between the two sounds. The user listens and decides.


Sound Effects Skill

Suno's Sounds makes short audio from a description — an effect, a one-shot or a seamless loop, not a song. One call makes two sounds of about 20 seconds and costs 1 song credit each (measured).

The two can differ in kind. Measured: asked for "soft rain on a window", one was rain and the other had a piano melody in it. Both get heard before either is used. Background: ${CLAUDE_PLUGIN_ROOT}/reference/suno/mcp-api-mapping.md § Sounds.

Step 1: Resolve

Expected format: <album-name> "<what the sound is>" [--loop] [--fit <track>] [--title "<name>"]

  • The description says what is heard, in a few plain words: "rain on a tin roof", "vinyl crackle", "a single wooden door knock", "boom bap drums". To keep music out of an effect, say so: "rain on a window, no music".
  • --loop asks for a seamless loop; without it the sound is a one-shot.
  • --fit <track> gives the sound that track's tempo and key. Use it for musical material: a drum loop, a bass riff, a riser. Leave it off for an effect or an ambience — a tempo and a key pull the result toward music.

Step 2: Check the Suno Server

Follow § Preflight in the reference doc: find sounds-create, audio-wait and account-usage, load their current definitions, and read account-usage. Not connected → say this needs Suno's MCP server and stop; on suno.com it is Sounds. Suno offers Sounds from the Pro plan up.

Step 3: Build the Request

get_sound_request(album_slug, description, loop, fit_track, title).

  • found: false, or a blocking problem → show it and stop.
  • nothing_to_fit (a warning) → the track names no tempo and no key. Say so and carry on unfitted.

Step 4: Show the Summary and Get a Go-Ahead

READY TO MAKE A SOUND — [Album]

  Sound      rain on a window, no music
  Kind       loop
  Fitted     96 BPM, G (from 03 Third Track)
  Makes      2 sounds of about 20 seconds
  Cost       2 song credits. You have 9,841.
  After      open both (free).

Go ahead?   yes · cancel

Leave out Fitted when nothing is fitted. Wait for an explicit yes. Silence is not a yes.

Step 5: Send It

  1. Fix an idempotency key before the first attempt: {album_slug}/sound/{UTC timestamp}.
  2. Call sounds-create with the keys of request and the idempotency_key. Add nothing.
  3. Two ids come back in clip_ids. Call audio-wait with both and until: "complete".
  4. sounds-create takes no workspace. If the tool's answer had a workspace, file the two there: workspaces-update-clips(id: <workspace.id>, add: [<both ids>]). Filing is free.
  5. Call account-usage again. Credits spent = balance before − balance after.

Step 6: Let the User Hear Them

Report both, and open their song pages with open_song_page, as /bitwize-music:listen opens takes — that is free:

SOUNDS — rain on a window (loop)

  1   0:20   https://suno.com/song/...
  2   0:20   https://suno.com/song/...

  Spent 2 song credits. 9,839 left.

If the user wonders what one of them holds, Suno's listening agent says for free: audio-listen(id, question: "What is heard in this clip? Is there any music or voice in it?"), then audio-listen-wait.

Step 7: Save One (on Request)

When the user wants a sound on disk, download it as /bitwize-music:import-audio § Importing From Suno step C does: audio-download(id, format: "wav") without confirm first. A requires_credit answer needs the user's yes, because it spends a download credit. Save it to the matching path in save.files — sounds/ under the album's audio folder, which the mastering tools never read.

Nothing in the plugin mixes a sound into a track. The file is for a DAW or a video edit.

Remember

  1. Cost first, then a go-ahead, then one call. Two sounds, two credits.
  2. Hear both. They can differ in kind.
  3. Fit only musical material to a track.
  4. You never choose between them.

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.