jupytext
Pass
Audited by Gen Agent Trust Hub on Mar 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill includes shell scripts (scripts/init_project.sh, scripts/sync_all.sh) designed to automate project directory setup and synchronize notebook files using the jupytext command-line tool.
- [REMOTE_CODE_EXECUTION]: The primary workflow facilitates the conversion and execution of code within notebooks using papermill. It also provides instructions for setting up various language kernels from trusted sources.
- [EXTERNAL_DOWNLOADS]: The skill references the official Jupytext GitHub repository for pre-commit hooks and utilizes standard registries like Conda-forge, PyPI, and CRAN for dependency management.
- [PROMPT_INJECTION]: The skill employs strong instructional framing (e.g., "IRON LAW") in SKILL.md to ensure the agent strictly follows a verification pipeline (convert then execute) before reporting success.
- [PROMPT_INJECTION]: Indirect injection surface analysis: Untrusted scripts/notebooks enter the context through file reads (Ingestion). Boundary markers consist of strict output verification requirements (Execution Enforcement). Capabilities include full code execution via Papermill. No specific input sanitization is performed as the skill is intended for technical users executing their own code.
Audit Metadata