The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests text from external URLs provided by the user, which is then processed by the agent to create new content and image prompts. There are no explicit boundary markers or sanitization steps to filter malicious instructions within the fetched content. Every Category 8 assessment confirms ingestion via fetch_article.py, no boundary markers for external data, capability inventory involving network requests to redfox.hk, and lack of content sanitization beyond basic HTML tag removal.\n- [EXTERNAL_DOWNLOADS]: The scripts fetch_article.py and generate_image.py download HTML and image content from remote URLs based on user input for article creation and style reference.\n- [DATA_EXFILTRATION]: The skill transmits content prompts and images to the vendor's API at redfox.hk. This is a legitimate vendor resource from the skill author (redfox-data) used for the skill's primary function.\n- [SAFE]: The web scraping and image download components disable SSL certificate verification (ssl.CERT_NONE), which is a security best-practice violation that could potentially expose connections to man-in-the-middle attacks.

visual-ops-writer