mj-xhs-note-batch-analysis

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes getnote CLI commands to perform operations like authentication checks, saving notes, and fetching content.\n- [DATA_EXFILTRATION]: The skill collects data from a user's logged-in Xiaohongshu session (note titles, URLs, and transcripts) and transfers it to the Get笔记 platform. The skill contains explicit instructions to avoid harvesting sensitive browser data such as cookies, passwords, or local storage.\n- [PROMPT_INJECTION]: The skill processes untrusted content from Xiaohongshu notes, creating an attack surface for indirect prompt injection where malicious instructions in the notes could affect the agent's synthesis.\n
  • Ingestion points: Note content is ingested via getnote note <id> --field web_content as described in SKILL.md.\n
  • Boundary markers: No specific delimiters are specified in the instructions to wrap the external content during processing.\n
  • Capability inventory: The skill uses the getnote CLI tool and a browser plugin for JavaScript execution in the page context.\n
  • Sanitization: No security-focused sanitization or escaping of the input data is performed before synthesis.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 07:51 AM
Security Audit — agent-trust-hub — mj-xhs-note-batch-analysis