The Agent Skills Directory

EXTERNAL_DOWNLOADS (MEDIUM): The downloader.py script uses urllib.request.urlopen to fetch content from arbitrary URLs provided as input. This can be exploited for Server-Side Request Forgery (SSRF) to access internal network services or cloud metadata endpoints.
DATA_EXFILTRATION (MEDIUM): The flomo.py utility sends a POST request containing article summaries to a user-specified api_url. If an attacker provides a malicious URL, this could be used to exfiltrate sensitive information processed by the agent.
PROMPT_INJECTION (LOW): The skill is highly susceptible to Indirect Prompt Injection.
Ingestion points: utils/downloader.py via fetch_html and download_image which retrieve content from external WeChat URLs.
Boundary markers: None. The article content is parsed into Markdown and returned to the agent context without delimiters or warnings to ignore embedded instructions.
Capability inventory: The skill possesses network GET/POST capabilities via urllib and local file write capabilities for image storage.
Sanitization: The Parser class in utils/parser.py performs HTML unescaping and regex-based cleaning for display purposes but does not sanitize content for malicious instructions directed at the LLM.
DATA_EXFILTRATION (SAFE): Image downloading in utils/downloader.py uses MD5 hashing of URLs for filenames, which effectively prevents path traversal attacks during file writes.

wechat-article-fetcher