jupyter-notebook
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill inserts user-provided titles into notebook files, creating a surface for instruction injection if the input is untrusted. 1. Ingestion points: The title argument in the new_notebook.py script. 2. Boundary markers: Absent in the markdown title cell. 3. Capability inventory: The script writes to the filesystem and the skill workflow includes notebook execution. 4. Sanitization: Filenames are slugified but markdown content is not escaped.
- [External Downloads] (LOW): The skill recommends the installation of standard packages such as jupyterlab and ipykernel.
- [Command Execution] (SAFE): The skill executes a bundled Python script to automate notebook generation from local templates using standard libraries.
Audit Metadata