The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external tools including kicad-cli, pdftoppm, and pdfinfo using the subprocess module. All executions use argument lists rather than raw shell strings, which prevents shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from KiCad design files and project documentation. Ingestion points: view.py reads schematic and board files, as well as DESIGN.md. Boundary markers: No explicit delimiters are used when passing parsed data back to the agent. Capability inventory: The script performs file system operations (reads) and executes localized CLI tools; it lacks network access. Sanitization: External data is filtered through regular expressions and cast to numeric types before being used in tool arguments or logic.
[DYNAMIC_EXECUTION]: The view.py script includes logic to dynamically add a sibling directory to the Python path to import a schematic helper library. This is a common practice for modular skills in a shared repository and is considered benign.

kicad-view