The Agent Skills Directory

[SAFE]: The skill performs legitimate embedded systems debugging operations as described. No malicious behaviors, obfuscation, or suspicious network activities were detected during the analysis.
[COMMAND_EXECUTION]: The skill invokes local openocd and arm-none-eabi-gdb binaries to communicate with hardware. These commands are executed using subprocess.run and subprocess.Popen with argument lists, which is a secure method that avoids shell injection vulnerabilities. The use of a raw command sub-command allows for flexibility in hardware interaction, which is a standard requirement for debugging tools.
[INDIRECT_PROMPT_INJECTION]: The skill exposes capabilities to write to device memory and execute raw debugger commands based on user input. While these are high-impact operations for the connected hardware, they are core functional requirements for a debugging utility. The skill provides an operation_mode configuration (e.g., mode 3 requires confirmation) to manage the risk of executing dangerous commands accidentally and uses structured JSON for output parsing to minimize the risk of malicious instructions hidden in tool output.

openocd