The Agent Skills Directory

[SAFE]: The skill operates entirely within its defined scope of notebook management. All tools and scripts are used for static analysis and file manipulation without executing untrusted notebook code or performing network operations.
[PROMPT_INJECTION]: The skill processes untrusted user-provided .ipynb files, which constitutes a surface for indirect prompt injection. This is inherent to the skill's primary purpose.
Ingestion points: The scripts/assess.py script parses .ipynb files using standard JSON libraries, and the agent reads these files via the Read tool.
Boundary markers: No explicit delimiters or specific warnings against embedded instructions are provided when processing the notebook content.
Capability inventory: The skill has access to file system tools (Read, Write, Glob, Grep) and a restricted execution environment for the assessment script.
Sanitization: Data is handled as structured JSON and analyzed using regular expressions; no evaluation or execution of the code cells occurs within the skill's logic.

notebook