xiaohongshu-content-benchmark

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shared utility scripts to execute external tools such as the postplus CLI for billing confirmations and python3 -m yt_dlp for potential video metadata handling. These executions are scoped to the skill's operational requirements and use standard child process spawning.
  • [DATA_EXPOSURE]: The skill interacts with the user's local PostPlus configuration files (e.g., config.json) to retrieve session tokens for authenticating with the vendor's cloud API. This is standard behavior for vendor-provided tools requiring cloud-based scraping services.
  • [EXTERNAL_DOWNLOADS]: Data is fetched from Xiaohongshu (Rednote) through a hosted collection bridge. The implementation uses a secure transport layer (HTTPS or local socket bridge) to communicate with the PostPlus Cloud infrastructure.
  • [PROMPT_INJECTION]: The skill processes untrusted content (titles and descriptions from Xiaohongshu posts). While this represents an indirect prompt injection attack surface, the risk is mitigated by the skill's focus on data normalization and ranking rather than executing instructions found within the scraped content.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 01:13 PM