# Run Python Tasks in a Remote Sandbox

Local environments may lack libraries or isolation for short Python tasks. This skill submits code to inference.sh and returns console or file outputs.

## Install

```bash
npx skillstore add infsh-skills/python-executor
```

## Metadata

- Status: approved
- Slug: infsh-skills-python-executor
- Skillstore revision: r2
- Version status: missing
- Tree hash: e7dfc159ab8ec1467df6a41d7d9453591d31715785fa4a066460349610403e6f
- Author: infsh-skills
- GitHub username: infsh-skills
- License: MIT
- Repository: https://github.com/infsh-skills/skills/tree/main/tools/utilities/python-executor/
- Ref: 4121de961d1b6f2ffca856260e239505c302452c
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network
- Quality score: 50
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/infsh-skills-python-executor
- Manifest: https://skillstore.pages.dev/api/skills/infsh-skills-python-executor/manifest

## Capabilities

- Executes Python 3.10 code through the infsh/python-executor application.
- Uses common data libraries including NumPy, Pandas, SciPy, Matplotlib, Seaborn, and Plotly.
- Supports HTTP requests, HTML parsing, web scraping, and browser automation libraries.
- Processes images, audio, video, 3D models, PDFs, and SVG graphics.
- Returns files saved under the outputs directory with console output.

## Use Cases

- Explore Data Quickly: Summarize datasets, calculate statistics, and generate charts without preparing a local Python environment.
- Automate Public Web Research: Request public pages or APIs, extract structured fields, and return cleaned files.
- Generate Media Assets: Create or transform images, videos, audio, documents, and 3D files with pre-installed libraries.

## Prompt Templates

### Run a Simple Calculation

```
Run Python that prints the first ten Fibonacci numbers. Return the console output and explain any execution error.
```

### Analyze a Dataset

```
Analyze {dataset} with pandas. Summarize missing values and key statistics, then save a chart to outputs/summary.png.
```

### Extract Public Web Data

```
Fetch {public_url}, extract {fields}, and save clean rows to outputs/results.csv. Respect access rules and report request failures.
```

### Build a Reproducible Media Pipeline

```
Process {input_asset} with {library}. Validate inputs, limit resources, save files under outputs, and report reproducible execution details.
```

## Limitations

- Requires the belt CLI, an inference.sh account, authentication, and service availability.
- Code and embedded data are submitted to a third-party remote service.
- The environment is CPU-only, uses Python 3.10, and limits memory and runtime.
- Interactive displays are unsupported; generated files must be saved under outputs.

## Best Practices

- Review generated Python and destination domains before authorizing remote execution.
- Keep secrets, personal data, and regulated information out of submitted code and inputs.
- Set realistic timeouts and save every required artifact under the outputs directory.

## Anti Patterns

- Do not submit credentials, private source code, or confidential datasets.
- Do not scrape websites without checking access rules, rate limits, and applicable terms.
- Do not expect GPU acceleration, interactive windows, persistent state, or unlimited runtime.

## Security Audit

- Audited at: 2026-08-06T12:03:29.143\+00:00
- Summary: Most static command findings are Markdown formatting false positives. Real belt and npx commands create external execution and supply-chain exposure, while the workflow sends Python code to inference.sh for remote processing.

## Stats

- Views: 97
- Downloads: 11
- Favorites: 0
- Popularity score: 0