The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill performs network requests to fetch article content and images from mp.weixin.qq.com. It also uses external resources like mmbiz.qpic.cn for image downloads.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external URLs which could contain malicious instructions designed to influence the behavior of the AI agent processing the content.
Ingestion points: Article body, titles, and metadata fetched from external WeChat URLs in scripts/fetch_wechat_article.py.
Boundary markers: No specific delimiters or warnings are used when the agent processes or saves the fetched content to prevent the agent from obeying embedded instructions.
Capability inventory: The skill has the capability to write files to the local system (pathlib.Path.write_text, pathlib.Path.write_bytes) and perform further network requests (requests.get).
Sanitization: Filenames are sanitized using safe_filename to prevent path traversal, but the actual content (HTML/Markdown) is not sanitized for potential prompt injection vectors before being presented to the agent.
[COMMAND_EXECUTION]: The skill is intended to be executed as a Python script via the command line, accepting URLs and directory paths as arguments, which are then used in network and filesystem operations.

wechat-article-fetcher