markdown-to-image
Pass
Audited by Gen Agent Trust Hub on Apr 8, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external Markdown content and renders it into HTML for image generation. The lack of robust sanitization on input data creates a surface where malicious formatting or embedded HTML tags could influence the output or the rendering environment.
- Ingestion points: Markdown files read via the
--fileargument inscripts/md2img.pyandscripts/md_to_image.py. - Boundary markers: No delimiters or warnings are used to isolate untrusted content during prompt interpolation or rendering.
- Capability inventory: The skill has file system access (read/write), network capabilities (via
requests), and browser automation control (viaplaywright). - Sanitization: The scripts use basic regular expressions to replace specific Markdown syntax with HTML tags, but do not perform comprehensive sanitization or escaping of the input content.
- [EXTERNAL_DOWNLOADS]: The scripts fetch external resources during execution.
- Downloads font styles from Google's public font service during the rendering process.
- Retrieves video thumbnails from YouTube's CDN (
i.ytimg.com) and metadata fromnoembed.com. - [COMMAND_EXECUTION]: The skill utilizes the
playwrightlibrary to launch and manage a Chromium browser instance to perform the conversion of HTML templates into PNG images.
Audit Metadata