The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables an indirect prompt injection surface by fetching full article content from external URLs and feeding it to the agent for summarization and analysis.
Ingestion points: The fetch_url_content function in scripts/fetch_news.py downloads and extracts text from URLs provided by news sources when the --deep flag is active.
Boundary markers: The instructions do not specify the use of delimiters or warnings for the agent to ignore commands that may be embedded in the scraped text.
Capability inventory: The skill performs network read operations using the requests library.
Sanitization: The script removes non-textual HTML elements (scripts, styles, navigation, footers) but does not perform semantic sanitization of the article text for malicious instructions.
[EXTERNAL_DOWNLOADS]: The skill fetches real-time news data from established technology and financial platforms including Hacker News, GitHub, Product Hunt, 36Kr, Tencent News, Wall Street CN, V2EX, and Weibo. When deep fetching is enabled, the script will also download content from arbitrary third-party URLs linked by these platforms.

news-aggregator-skill