The Agent Skills Directory

[CREDENTIALS_UNSAFE]: Hardcoded authentication token detected in multiple files.
Evidence: The token 2f9f88dbb743423dbf0a8db2977c49eb is hardcoded in references/api-spec.md and scripts/fetch_hot_articles.py for API access.
[COMMAND_EXECUTION]: The skill executes local Python scripts that utilize insecure network configurations.
Evidence: scripts/fetch_hot_articles.py and scripts/generate_hot_html.py are executed during the workflow. The script fetch_hot_articles.py explicitly disables SSL certificate verification and hostname checking, which is a significant security risk exposing the connection to Man-in-the-Middle (MitM) attacks.
[EXTERNAL_DOWNLOADS]: The skill references a third-party library from a well-known service for HTML to PDF conversion.
Evidence: scripts/generate_hot_html.py includes a reference to html2pdf.js hosted on the cdnjs.cloudflare.com CDN.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing external content.
Ingestion points: Article data (titles, summaries) is fetched from a remote API by scripts/fetch_hot_articles.py.
Boundary markers: Absent. The skill does not use delimiters or instructions to ignore commands within the fetched article content.
Capability inventory: The skill can execute shell commands and write to the local file system.
Sanitization: Absent. Fetched content is displayed without validation or escaping.

wechat-10w-hot