The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and analyzes untrusted codebase files (UI controllers, README files, and metadata) to derive marketing benefits. Instructions or malicious text hidden in these files could influence the subsequent image generation prompts.
Ingestion points: Project codebase analysis (UI files, models, and metadata) in SKILL.md during the Benefit Discovery phase.
Boundary markers: None; the skill does not use specific delimiters or instructions to prevent derived text from overriding generation parameters.
Capability inventory: Local execution of Python scripts, system command execution via bash (sips utility), and AI image generation capabilities via edit_image and generate_image tools.
Sanitization: None; codebase-derived strings are interpolated directly into prompts for the AI image tools without validation or escaping.
[COMMAND_EXECUTION]: The skill utilizes the bash tool to execute internal Python processing scripts (compose.py, showcase.py) and the macOS sips (Scriptable Image Processing System) utility. These tools are used locally to composite images, crop them to specific aspect ratios, and resize them for App Store Connect compatibility.

aso-appstore-screenshots