xiaohongshu-search

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a provided local Python script (scripts/fetch_xhs_hot_articles.py) to query trending data. This is a controlled execution of the skill's own code.
  • [EXTERNAL_DOWNLOADS]: The Python script performs network requests to redfox.hk to fetch note metadata. This domain is the official API endpoint for the skill's vendor (Redfox Hub) and is used exclusively for the intended search functionality.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external content from Xiaohongshu notes, which represents an injection surface. However, the risk is mitigated by strict output formatting requirements in SKILL.md (e.g., Markdown tables) and character escaping for HTML generation in the Python script.
  • Ingestion points: Data is fetched from https://redfox.hk/story/api/xhs/search/search in scripts/fetch_xhs_hot_articles.py.
  • Boundary markers: The agent is forced to use specific table templates and JSON structures, which act as delimiters.
  • Capability inventory: The script performs network calls to a specific vendor API and writes a local HTML report.
  • Sanitization: The script performs HTML entity escaping on note titles and descriptions before writing to the report.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 03:02 PM