lovstudio-any2pdf
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The script downloads emoji assets from the Cloudflare CDN (cdnjs.cloudflare.com), which is a well-known and trusted service.
- [EXTERNAL_DOWNLOADS]: The script fetches remote images from arbitrary URLs provided within the user-supplied markdown documents using the urllib.request library.
- [COMMAND_EXECUTION]: The agent executes the md2pdf.py Python script to process markdown files and generate PDF documents.
- [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection as it ingests untrusted markdown data. The agent is specifically instructed in SKILL.md to read parts of the input document (title and first paragraphs) to inform the generation of themed images.
- Ingestion points: The agent reads the title and opening paragraphs of the user-provided markdown file (scripts/md2pdf.py).
- Boundary markers: There are no delimiters or warnings to ignore embedded instructions within the processed document.
- Capability inventory: The skill can execute local scripts and perform network operations.
- Sanitization: Content is escaped for PDF rendering but not sanitized for prompt-based injection attacks.
Audit Metadata