The Agent Skills Directory

[SAFE]: The skill operates as a local text-processing utility. It analyzes and modifies C/C++ source code files based on user-defined naming conventions using standard Python libraries.\n- [PROMPT_INJECTION]: The skill processes external, untrusted source code downloaded from the internet. While this creates a surface for indirect prompt injection (e.g., malicious instructions in code comments), the risk is mitigated because the tool does not possess capabilities for network communication or arbitrary command execution. The output of the tool is intended for integration into a firmware project, where standard compilation and review processes apply.\n
Ingestion points: scripts/bsp_adapter.py reads files from user-specified directories via the --scan and --adapt arguments.\n
Boundary markers: Absent; the script reads raw file content for regex analysis.\n
Capability inventory: File system write access (pathlib.Path.write_text in scripts/bsp_adapter.py) to generate adapted source files. No network or subprocess capabilities.\n
Sanitization: Absent; the tool performs regex replacements on raw text.

peripheral-driver