xiaohongshu-recruiter
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local scripts to perform its tasks. The script
scripts/publish_xiaohongshu.pyutilizessubprocess.Popento launch a standalone instance of Google Chrome with remote debugging flags. This design is intended to allow the browser to remain open for user interaction (such as manual login) and to persist after the automation script finishes. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided job descriptions to generate public social media posts. While this is the intended functionality, it creates a surface for untrusted data to be processed and published.
- Ingestion points: User-supplied job details passed via CLI arguments to the automation scripts.
- Boundary markers: None; the data is directly interpolated into the browser automation workflow.
- Capability inventory: The skill can generate images locally using Node-Canvas and control a web browser via Playwright to perform file uploads and form filling on the Xiaohongshu platform.
- Sanitization: The script performs basic length validation (truncation to 20 characters) for titles, but does not sanitize the body text against potential prompt-based manipulation within the context of the recruitment post.
Audit Metadata