The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of several standard Python libraries, including google-genai, Pillow, and playwright, which are necessary for image processing and rendering functionality.\n- [COMMAND_EXECUTION]: The skill executes bundled Python scripts (gen_image.py, profile_store.py, render_to_png.py) to perform its primary tasks. These scripts are managed as internal vendor resources.\n- [DYNAMIC_EXECUTION]: The render_to_png.py script utilizes the Playwright library to launch a Chromium browser instance to render HTML and SVG content into static images.\n- [PERSISTENCE]: The skill maintains a persistent user profile in a local JSON file via the profile_store.py script, allowing it to remember brand assets and user preferences across different sessions.\n- [INDIRECT_PROMPT_INJECTION]: The tool processes HTML and SVG files for rendering, creating a potential surface for indirect prompt injection if those files contain malicious instructions.\n
Ingestion points: The render_to_png.py script reads HTML documents from the local project workspace for rendering.\n
Boundary markers: The SKILL.md workflow requires the agent to explicitly identify project files and respect project-level output directories before execution.\n
Capability inventory: The skill has the capability to perform local file system writes and execute automated browser rendering via Playwright.\n
Sanitization: The rendering script processes HTML files directly from the local workspace without additional sanitization filters.

lov-image-creator