The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by performing a comprehensive analysis of the project's codebase and reading external simulator screenshots. This data is used to identify app benefits and formulate prompts for AI image generation. The skill does not implement explicit boundary markers or sanitization logic to prevent malicious instructions embedded within the code or image metadata from influencing the agent's behavior during the generation phase.
[COMMAND_EXECUTION]: The workflow relies on the generation and execution of shell commands. This includes running local Python scripts (compose.py, showcase.py) and the macOS system tool sips for image cropping and resizing. These commands interpolate variables such as file paths, background colors, and benefit headlines derived from both user input and automated codebase analysis.
[EXTERNAL_DOWNLOADS]: The installation instructions require the user to download and install the Pillow library via pip and a third-party Model Context Protocol (MCP) server (@houtini/gemini-mcp) from NPM. These external components are necessary for the skill's core functionality of image composition and AI-driven visual enhancement.

aso-appstore-screenshots