The Agent Skills Directory

[PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection, where malicious instructions in source material could hijack the image generation process.
Ingestion points: scripts/create_slides.py reads user-provided text through the load_prompts and outline_to_prompts functions.
Boundary markers: Absent. The skill uses simple f-strings to build prompts, providing no separation between system instructions and untrusted data.
Capability inventory: subprocess.run is used to execute external Python scripts from the nano-banana skill.
Sanitization: Absent. No escaping, validation, or filtering is performed on input text before it is passed as a command-line argument.
[COMMAND_EXECUTION] (MEDIUM): The script executes local files via subprocess.run using paths constructed from the user's home directory.
Evidence: scripts/create_slides.py targets ~/.claude/skills/nano-banana/scripts/generate_image.py using Path.home().
Risk: While using an argument list prevents direct shell injection, the skill relies on the presence and security of scripts outside of its own package.

slide-creator