The Agent Skills Directory

[SAFE]: The skill operates as a thin orchestration layer for embedded development tools. No malicious behavior, obfuscation, or data exfiltration attempts were identified. All activities are consistent with the skill's stated purpose of coordinating build and flash workflows.
[COMMAND_EXECUTION]: Subprocess execution is used to orchestrate sibling scripts (e.g., 'keil_build.py', 'gcc_build.py', 'jlink_exec.py'). These calls are implemented using subprocess.run with argument lists and without shell=True, which is a secure practice that prevents shell injection. All executed paths are restricted to the skill's known internal directory structure.
[DATA_EXFILTRATION]: File access is localized to reading and writing configuration ('config.json') and state ('state.json') within the workspace's '.embeddedskills' directory. No network operations or access to sensitive global system files (such as SSH or AWS credentials) were detected.

workflow