The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/run_lovpen_cli.py is used to invoke the lovpen-cli backend. The implementation uses the Python subprocess.run function with shell=False, which correctly avoids shell injection risks by passing arguments as an array.
[EXTERNAL_DOWNLOADS]: The skill documented requirements for external tools such as Node.js, Python, and a Chromium-based browser. It also relies on the local installation of the lovpen-cli package to perform its primary rendering tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes external Markdown data provided by the user, which presents a surface for indirect prompt injection.
Ingestion points: Markdown input file content is processed by the renderer in scripts/run_lovpen_cli.py via the render command.
Boundary markers: There are no explicit boundary markers or "ignore instructions" delimiters used when passing the file content or paths to the CLI.
Capability inventory: The skill has the capability to read and write to the local filesystem and execute the lovpen-cli subprocess.
Sanitization: Input parameters are handled as distinct arguments in a list, preventing command-string manipulation at the shell level.

lovpen-cli