The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/serial_monitor.py uses subprocess.run to execute openocd and allows arbitrary commands to be passed via the --openocd-command flag. This could be abused if an agent populates this argument using untrusted input.
[PROMPT_INJECTION]: The skill processes serial data from external hardware, creating an indirect prompt injection surface. Ingestion points: Serial port input in scripts/serial_monitor.py. Boundary markers: Absent; the agent is not provided with delimiters to distinguish data from instructions. Capability inventory: Subprocess execution of openocd and file writing. Sanitization: No validation or sanitization is performed on the serial data.
[EXTERNAL_DOWNLOADS]: The skill documentation directs users to install the pyserial library and the openocd system utility using standard package managers.

serial-monitor