The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs ripgrep (rg) to conduct a deep search across the entire $HOME directory, including hidden and ignored files, to find KiCad binaries. While the search is filtered by specific filename globs (e.g., *kicad*.appimage), the use of --no-ignore and --hidden on the entire home directory is an aggressive discovery pattern that exposes local directory structures to the agent.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external design data (KiCad schematics and PCBs) while maintaining access to powerful tools like the KiCad CLI and Python script execution.
Ingestion points: KiCad project files (.kicad_sch, .kicad_pcb, .kicad_pro), symbol and footprint library tables, and local project scripts.
Boundary markers: Absent. The instructions do not define delimiters or provide guidance on ignoring natural language instructions that might be embedded in CAD comments, net names, or metadata.
Capability inventory: Execution of kicad-cli subcommands (ERC, DRC, exports), generation and execution of Python automation scripts using kicad-python (kipy) and kicad-sch-api, and system GUI automation.
Sanitization: Absent. There is no requirement for the agent to sanitize or validate strings retrieved from CAD files before using them in further logic or command generation.
[DYNAMIC_EXECUTION]: The workflow encourages the agent to generate and maintain 'durable' Python scripts for complex PCB and schematic transformations. This involves the runtime creation and execution of code within the project environment using the kicad-python and kicad-sch-api libraries.

kicad-workflow