The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/convert.py script executes the pandoc binary using subprocess.run() with a list of arguments. This method avoids shell interpretation and protects against command injection vulnerabilities.
[DATA_EXFILTRATION]: The skill is designed to read and write document files for conversion. It does not attempt to access sensitive system directories, environment secrets, or credentials, nor does it perform unauthorized network operations.
[PROMPT_INJECTION]: The skill processes external documents which serves as an ingestion point for untrusted data, characterizing an indirect prompt injection surface.
Ingestion points: scripts/convert.py (via the input argument).
Boundary markers: Not explicitly present in the script.
Capability inventory: subprocess.run in scripts/convert.py.
Sanitization: Relies on the underlying Pandoc tool's parsing logic.

pandoc