Guide · Agent skill
How to use the Krita Command-Line Control agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The cli-anything-krita skill teaches an AI agent to drive the Krita digital painting application through a command-line interface instead of its graphical window. It is aimed at artists, studios and developers who want to batch-process or batch-export Krita files, or have an agent build a layer stack and export a finished illustration. After this guide you will know what the skill does, how to install it, how to phrase requests so the agent uses it, and what safety points to check first.
Krita Command-Line ControlHKUDS/CLI-Anything · Apache-2.0 Download the skill Quick start
- 1Confirm Krita and Python are installed. Check that Krita is installed on the system and that Python is version 3.10 or newer.
- 2Download the skill ZIP. Download the cli-anything-krita ZIP from this page.
- 3Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder.
- 4Install the command-line tool. From the krita/agent-harness folder of the HKUDS/CLI-Anything repository, run pip install -e . as shown in the source.
- 5Ask for a Krita task in plain language. Describe the file, layers, filter or export you want, mentioning Krita.
- 6Review destructive commands. Check any export with overwrite or layer removal before the agent runs it, and keep backups of originals.
What the skill lets the agent do
The skill is a command-line harness for Krita. Once it is available, the agent can create project files, add, remove and edit layers, apply image filters, resize the canvas, and export renders. The source lists export presets including PNG, JPEG, PSD, PDF, SVG, WebP, GIF, TIFF and BMP.
It also supports undo and redo through a session history, and every command accepts a --json flag so the agent can read machine-readable output.
- Project management: create, open, save and inspect a project file
- Layers: add, remove, list and set properties such as opacity, visibility and blending mode
- Filters: blur, sharpen, levels, curves, hue-saturation and others
- Canvas: resize by pixel dimensions or by resolution
- Export: render single images or export animation frames
Install the skill
You can install the skill by downloading the ZIP from this page and unzipping it into one of two folders. Use ~/.claude/skills/ for a personal skill available across your projects, or a project's .claude/skills/ folder to limit it to one project. Alternatively, in the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
The skill's SKILL.md notes that the command-line tool itself must be installed separately. Its install step is run from the krita/agent-harness folder of the HKUDS/CLI-Anything repository with pip install -e ., so you need Krita installed on the system and Python 3.10 or newer.
- Krita must be installed on the same system
- Python 3.10 or newer is required
- Run the pip install from the krita/agent-harness folder of the repository
cd krita/agent-harness && pip install -e .Trigger the skill with natural requests
You do not need to type commands yourself. Ask the agent for a result in plain language that mentions Krita, layers, filters, batch processing or exporting, and the skill's description tells the agent when it applies. Name the file and the outcome you want, such as the project file, the layer names, the filter and the export format.
Be specific about destructive actions. If you want the agent to overwrite an existing export or remove a layer, say so, because those commands change files without confirmation.
- "Create a 3000 by 4000 Krita project called Illustration and save it as art.json."
- "In art.json, add a layer called Sketch at 180 opacity, then export final.png."
- "Apply a blur to the Background layer in art.json and export it as a JPEG."
Example: build a layer stack and export
A typical job creates a project, sets up named layers, applies an effect to one layer and exports the result. The example below is the workflow from the skill's source, using JSON output for the agent to read.
The final export command includes --overwrite, which replaces an existing final.png without asking.
cli-anything-krita --json project new -n "Illustration" -w 3000 -h 4000 -o art.json
cli-anything-krita --json -p art.json layer add "Background" -t paintlayer
cli-anything-krita --json -p art.json layer add "Sketch" -t paintlayer --opacity 180
cli-anything-krita --json -p art.json filter apply blur -l "Background"
cli-anything-krita --json -p art.json export render final.png -p png --overwriteSafety notes: scripts and files
The skill runs scripts, so treat it like any other code you install. The skill's listed requirements say it does not use the network. It does, however, drive your local Krita application and change project and export files.
Work on copies of important files first. Commands such as export with --overwrite and layer removal modify output without confirmation, so review what the agent plans to run before approving it.
- Back up original .kra or project files before batch runs
- Check the target output path before using --overwrite
- Review layer remove commands before the agent runs them
Tips for reliable results
Use the --json flag when an agent is reading output, and use session history to check or undo recent steps. Name layers consistently, since layer commands refer to layers by name. Run cli-anything-krita status to check the harness state before a long job.
List the available filters and export presets with the filter list and export presets commands, so you can ask for names the tool actually recognises.
- Name layers uniquely so commands like layer set target the right one
- Use export formats and export presets to confirm supported options
- Start with a small test project before running a large batch
cli-anything-krita export presets
cli-anything-krita export formats
cli-anything-krita statusWhat people use it for
- Creating a layered project file with named layers for a sketch, colours and effects
- Batch-applying a filter such as blur or levels to a layer across many projects
- Exporting a finished illustration to PNG, JPEG, PSD or PDF from a script or agent
- Exporting animation frames from a folder of Krita frames to PNG
- Resizing canvases to a new pixel size or resolution in an automated pipeline
Frequently asked questions
Do I need Krita installed to use this skill?
Yes. The skill drives the local Krita application, and the source lists Krita installed on the system as a requirement. Without Krita the commands will not work.
What do I need to install cli-anything-krita?
You need Krita, Python 3.10 or newer, and the tool installed from the krita/agent-harness folder of the HKUDS/CLI-Anything repository with pip install -e .. The skill itself is installed by unzipping it into a skills folder.
Is the Krita command-line skill free to use?
The skill is licensed under Apache-2.0, as listed on this page. Check the license file in the repository for the full terms.
Does the skill need an internet connection?
The listed requirements say it does not use the network. It runs local scripts that work with your installed Krita application.
Can it paint or draw for me?
No. The skill covers project structure, layers, filters and export. It is not meant for brush work or freeform painting.
How is this different from using Krita's normal interface?
The skill lets an agent operate Krita without a graphical interface, which suits repeated steps across many files. For one-off hand-painted work, the normal Krita window is the better tool.
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 artists and studios who need to batch-process or batch-export Krita files, and developers building image pipelines around Krita. It is a poor fit if you do not have Krita installed, want help with brush painting, or need to protect files from automatic overwrites and layer removal.
Download the skill