python-executor — Claude Code skill: what it does, how to install
Runs Python 3.10 scripts inside an isolated inference.sh sandbox. Helpful when your agent needs to crunch data, scrape websites, or render graphics without touching your local system.
npx skills add 101-skills/superpowers --skill python-executor -a claude-codeWhat python-executor does
The skill offloads script execution to an inference.sh container via the belt CLI. The sandbox runs Python 3.10 with an execution window between 1 and 300 seconds, defaulting to 30 seconds. The base configuration provides 8GB RAM, while memory-heavy jobs can switch to a 16GB tier.
Over one hundred libraries come pre-installed. The environment bundles data tools like pandas and scipy, browser automation drivers like playwright and selenium, alongside media packages like moviepy and trimesh. Every artifact written to the outputs/ directory is captured and returned directly to the caller.
This container provides CPU compute only, meaning GPU model training will fail. Interactive GUI displays are unsupported, so scripts must save plots via plt.savefig() instead of plt.show(). Before executing scripts, you need the belt CLI installed and authenticated with belt login.
What it helps with
- scrape web pages using Playwright or BeautifulSoup
- generate charts with Matplotlib and export images
- inspect and transform 3D STL meshes with Trimesh
- render short video clips with text overlays via MoviePy
Example prompts for the agent
- Run this scraping script using python-executor and save the titles
- Analyze the sales dataset with Pandas and save a chart to outputs/chart.png
- Generate an icosphere STL file using Trimesh and return the output
How to install python-executor
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill python-executor -a claude-code2Manually: put the files in place
~/.claude/skills/python-executor/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/python-executor/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill python-executor -a codex2Manually: put the files in place
.agents/skills/python-executor/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill python-executor -a cursor2Manually: put the files in place
.cursor/skills/python-executor/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install 101-skills/superpowers python-executor2Manually: put the files in place
.github/skills/python-executor/SKILL.md1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill python-executor -a gemini-cli2Platform CLI
gemini skills install https://github.com/101-skills/superpowers --path tools/utilities/python-executorManually: put the files in place
.devin/skills/python-executor/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill python-executor -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about python-executor
How to install python-executor in Claude Code?
Run npx skills add 101-skills/superpowers --skill python-executor -a claude-code. You also need the belt CLI installed and authenticated via belt login.
What does python-executor do?
It executes Python code in a secure cloud sandbox loaded with Pandas, Playwright, OpenCV, and other libraries. Output files saved in outputs/ are passed back automatically.
What are the environment limits of python-executor?
It runs strictly on CPU with no GPU support. Memory defaults to 8GB with an optional 16GB mode, and runtime is bounded between 1 and 300 seconds.
