The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes content from untrusted .ipynb files provided by the user. Reading and parsing external notebook content into the agent context creates an indirect prompt injection surface where instructions embedded in the notebook cells could influence the agent's behavior.
Ingestion points: The skill reads notebook.ipynb using json.load in Step 1 and processes cell source code in multiple steps.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the processing logic.
Capability inventory: The skill utilizes the Write tool to create new files (utils.py, requirements.txt, script.py) and the Bash tool to execute external commands.
Sanitization: Content is filtered using regex (e.g., Step 5 for imports) and standard conversion libraries (nbconvert), but raw cell strings are otherwise handled directly.
[EXTERNAL_DOWNLOADS]: The skill instructions suggest the installation of well-known third-party development tools using pip to facilitate notebook maintenance.
Evidence: References to black[jupyter], nbqa, and jupytext are provided as automated cleanup tools.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run standard utilities for notebook conversion, formatting, and linting.
Evidence: Use of jupyter nbconvert, black, and nbqa commands to process local files.

jupyter-notebook-assistant