The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided topics, files, or conversation summaries to generate infographic content. This data is interpolated into HTML files which are then rendered by a headless browser (Playwright). A malicious input could potentially attempt to inject HTML tags or scripts to influence the rendering environment.
Ingestion points: Target argument, conversation summary, topic, file, skill, or workflow (referenced in SKILL.md).
Boundary markers: Absent; no specific instructions for escaping or delimiting user content within the generated HTML.
Capability inventory: Python script execution (scripts/render_infographics.py), Playwright browser rendering (page.goto), and file creation/writing in the output/ directory.
Sanitization: Absent; the skill relies on the agent's generation logic without explicit sanitization steps for the resulting HTML.
[COMMAND_EXECUTION]: The skill uses a local Python script to perform rendering, which is executed as part of the infographic creation workflow.
Evidence: SKILL.md references the use of scripts/render_infographics.py for image generation.
[DYNAMIC_EXECUTION]: The skill generates content (HTML/CSS) at runtime which is subsequently processed and rendered by the Playwright engine.
Evidence: The rendering procedure describes creating independent HTML files per pattern and rendering them with Playwright.

visualize-as-infographic