The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run across multiple scripts to perform deterministic processing tasks.
scripts/render_animation_videos.py executes ffmpeg to generate preview videos.
scripts/finalize_pet_run.py chains several internal Python scripts using sys.executable.
scripts/generate_pet_images.py uses curl to interact with the OpenAI Image API (api.openai.com) when the primary $imagegen skill is unavailable. These executions are functional and restricted to the skill's operating directory or trusted services.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface in the way it handles user-controlled text.
Ingestion points: scripts/prepare_pet_run.py accepts user-supplied strings via the --pet-name, --description, and --pet-notes arguments.
Boundary markers: There are no delimiters or 'ignore' instructions surrounding the interpolated user data in the generated prompt templates.
Capability inventory: The generated prompt files (e.g., prompts/base-pet.md and row-specific prompts) are passed directly to the $imagegen skill or the OpenAI API to govern visual output.
Sanitization: No escaping or validation is performed on the user inputs before they are written to the prompt files. An attacker could provide a pet name containing instructions designed to override the visual style or bypass safety filters of the downstream image generator.

hatch-pet