The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/iar_builder.py executes the iarbuild.exe utility using subprocess.run. The implementation is secure as it passes arguments as a list and does not utilize a shell (shell=True), effectively preventing command injection.
[PROMPT_INJECTION]: The skill processes untrusted external data in the form of .ewp XML project files (ingestion point: scripts/iar_builder.py). While this presents a surface for indirect prompt injection via build logs or parsed configuration, the script uses the standard xml.etree.ElementTree for parsing and does not expose sensitive capabilities to the data being processed.
[DATA_EXFILTRATION]: No network operations or unauthorized file access patterns were detected. The skill only interacts with local build tools and project files consistent with its stated purpose.

build-iar