The Agent Skills Directory

[COMMAND_EXECUTION]: The script invokes external debugging utilities (gdb, nm, readelf) via subprocess.run. All calls use safe list-based arguments instead of a shell environment (shell=True), which mitigates shell command injection risks. Additionally, target ELF paths are validated using Path.exists() before processing, protecting against path-based GDB command scripting injections.
[SAFE]: The skill executes entirely within the local context, communicating only with a local GDB server (localhost). No data exfiltration, remote code downloads, or obfuscation patterns were detected.

rtos-debug