Skip to content
Get packs

CLI-Anything for Reasonix

Guides Reasonix to build, refine, test and validate CLI harnesses for GUI apps or repos

Source Download skill
Repository
HKUDS/CLI-Anything
Stars
58,084
Licence
Apache-2.0
Files
4 files

What it does

This skill teaches the Reasonix coding agent to act as a builder of CLI-Anything harnesses: small Python command-line wrappers that let an AI agent control a graphical application or source repository through a text interface. Once installed, asking Reasonix to make a CLI for a program such as GIMP makes it follow a multi-phase method (analyse the source, design commands, implement a Click CLI with JSON output and a REPL, write tests, and package it with pip). It may fetch the project's full methodology file from GitHub when it is not available locally.

Great for

  • Developers who want an agent to generate a command-line interface for a GUI tool or codebase
  • Teams already using Reasonix who want a repeatable harness-building workflow
  • Maintainers extending existing CLI-Anything harnesses with new commands

Skip it if

  • Non-developers looking for an everyday productivity tool: the output is Python packaging and test code that needs review
  • Users who need to wrap software they do not have rights to automate, since the skill generates code that drives real applications
  • Environments where an agent running unrestricted shell commands and pip installs is not acceptable, because the build runs bash, pip install -e and pytest

Try asking

  • “Build a CLI-Anything harness for ./gimp and produce the agent-harness folder with setup.py”
  • “Refine the existing harness in ./blender/agent-harness by adding missing commands without removing current ones”
  • “Write and run test_core.py and test_full_e2e.py for my audacity harness”
  • “Validate that my harness has a working cli-anything-<software> entry point and matching SKILL.md files”
Step-by-step guide: how to use CLI-Anything for ReasonixInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 4 original files, the Apache-2.0 licence and an attribution note.

    reasonix-skill.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip reasonix-skill.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip reasonix-skill.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Set agent.max_steps to 0 or at least 64 in reasonix.toml, since a full build needs 25-40 tool rounds and a small finite budget can truncate it
  • Review generated code and run the tests yourself before relying on the harness, especially the shell commands it executes
  • Install with scripts/install.sh, which copies the skill to ~/.reasonix/skills/cli-anything and refuses to overwrite an existing copy

What's inside

  • agents/reasonix.yaml
  • README.md
  • scripts/install.sh
  • SKILL.md

Redistributed unmodified from HKUDS/CLI-Anything under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.