The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/rtos_debugger.py uses the subprocess module to execute system commands including GDB (arm-none-eabi-gdb or gdb-multiarch), nm, and readelf. It also generates temporary GDB command scripts to facilitate batch analysis. These operations are necessary for the skill's primary function of firmware analysis.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect injection as it processes untrusted local files and command outputs.
Ingestion points: The --elf file path argument and the output of binutils/GDB are processed by scripts/rtos_debugger.py.
Boundary markers: None identified; file paths are interpolated directly into GDB scripts.
Capability inventory: The skill can execute commands via subprocess calls to local debugging tools.
Sanitization: The script performs basic path normalization (replacing backslashes) but does not validate inputs against newline injection, which could theoretically allow command injection within the generated GDB script environment. This is considered a low-risk implementation detail rather than a malicious vector in a development tool context.

rtos-debug