Guide · Agent skill
How to use the Audacity CLI agent skill (cli-anything-audacity)
Updated · written with AI from the project's own documentation and checked against our catalogue data
This guide explains how to add the cli-anything-audacity skill to an AI agent so it can edit audio through the command line instead of the Audacity window. It is written for readers who have never used agent skills. By the end you will know what the skill lets the agent do, how to install it, how to phrase requests that trigger it, and what to check before running it on private audio.
Audacity Audio Editing CLIHKUDS/CLI-Anything · Apache-2.0 Download the skill Quick start
- 1Install prerequisites. Install Python 3.10 or newer and Audacity on your system.
- 2Install the tool. Run pip install cli-anything-audacity.
- 3Download and unzip the skill. Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder.
- 4Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 5Verify the command. Run cli-anything-audacity --help to confirm the tool responds.
- 6Ask for a task with full paths. Describe the audio task and give absolute file paths for inputs and outputs.
- 7Check the output. Confirm the rendered file exists and review the JSON or history output.
What is an agent skill and what does this one do?
An agent skill is a folder of instructions (here, a SKILL.md file) that teaches an AI agent how to use a tool it might not know about. This skill documents cli-anything-audacity, a Python command-line interface for Audacity, a free audio editor.
Once installed, the agent can create project files, add and trim audio clips, adjust track volume and effects, add labels, and render the result to an audio file. Each command is stateful and supports undo and redo. The skill itself is documentation; the work is done by the cli-anything-audacity Python tool that the agent runs.
Install the Python tool and Audacity
The skill needs the underlying tool to be present. The source lists these prerequisites: Python 3.10 or newer, Audacity installed on your system, and the pip package cli-anything-audacity. Without Audacity and Python the skill will fail.
Install the package with pip:
- Python 3.10+
- Audacity installed on the system
- pip package cli-anything-audacity
pip install cli-anything-audacityInstall the skill into your agent
Download the skill ZIP from this page and unzip it. For a personal setup, place the unzipped skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
- Personal: unzip into ~/.claude/skills/
- Project only: unzip into the project's .claude/skills/
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Check that the command works
Before asking the agent to edit anything, run the help command to confirm the tool is installed. Starting the tool with no subcommand opens an interactive REPL with tab-completion and history.
cli-anything-audacity --helpHow to trigger the skill with natural requests
You do not need special syntax. Ask for an audio task that involves Audacity or the command line, and name the files and outcome you want. The source emphasises that the agent should use absolute paths for file operations, so including full file locations in your request helps.
Good requests name the input audio, the project file, and the output. For example: create a project, import two clips, trim the first one, and render the result to an audio file in a given folder.
- Make a new Audacity project at /full/path/podcast.json and add intro.wav to track one.
- Trim 2 seconds off the start of the clip on track 1 and undo that if the result sounds wrong.
- Add a label at 30 seconds reading 'Chapter 2' to the project.
- Render the project to a WAV file in my output folder and confirm the file exists.
Command areas the agent can use
The CLI is organised into groups. Knowing them helps you judge what is possible and read the agent's output.
- project: new, open, save, info, settings, json
- track: add, remove, list, set (name, mute, solo, volume, pan)
- clip: import, add, remove, trim, split, move, list
- effect: list-available, info, add, remove, set, list
- selection: set, all, none, info
- label: add, remove, list
- media: probe, check
- export: presets, preset-info, render
- session: status, undo, redo, history
Example workflow
A typical run creates a project, saves it as JSON, and renders it. The agent can request JSON output with the --json flag, which the source says is intended for agent consumption. Undo and redo cover up to 50 levels of history.
- Create the project file
- Add tracks and clips
- Apply volume or effect changes
- Check the result with info or history
- Render to an audio file and verify it exists
cli-anything-audacity --json project new -o myproject.jsonSafety notes: scripts and file access
This skill runs scripts. It does not use the network, according to its listing. The tool reads and writes local audio and project files, so check which paths the agent will touch before running it on private recordings. Render commands write output files, and the source's export example uses an --overwrite flag, so confirm the output path first to avoid replacing an existing file.
The source also recommends checking return codes (0 means success) and reading stderr when a command fails.
- Confirm input and output paths before any render or overwrite.
- Keep private recordings in a folder you have reviewed.
- Use the undo command if an edit goes wrong.
What people use it for
- Assembling several recorded clips into one track with trims and labels for a podcast episode.
- Batch-building projects from a list of audio files using repeatable commands.
- Adjusting track volume and pan across a set of tracks, then rendering the mix.
- Adding labels at timestamps to mark sections of a long recording.
- Scripting repeatable edits in a developer workflow that reads JSON output from each step.
Frequently asked questions
Is cli-anything-audacity free?
The skill's listing is licensed under Apache-2.0, which you can check in the repository. Audacity itself is described in the listing as a free audio editor. Confirm the licence terms in the repository before commercial use.
What do I need installed to use this skill?
You need Python 3.10 or newer, Audacity installed on your system, and the pip package cli-anything-audacity. The skill will fail without Audacity and Python.
How do I install the skill in Claude?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use or a project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
Does the skill need an internet connection?
The listing says it does not use the network. It does run scripts and reads and writes local audio and project files.
Can I undo mistakes made by the agent?
Yes. The CLI supports undo and redo with up to 50 levels of history, and you can view history with the session history command.
How does this compare with using Audacity directly?
This is a command-line wrapper with a limited command set, not a full visual editor. Pick it for scripted, repeatable edits. If you need the full range of Audacity's interface and tools, use the Audacity application itself.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits people who want an AI agent to batch-edit or assemble audio from the command line, and developers who want JSON output from each step. It is a limited wrapper, so people who need a full visual editor, or who work with sensitive recordings without checking file paths first, should choose another approach.
Download the skill