legal-case-analysis

Pass

Audited by Gen Agent Trust Hub on Aug 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The section titled '⛔ 使用授權鐵律' (Usage Authorization Iron Rule) uses language patterns often associated with prompt injection and behavioral locking. It employs phrases such as 'Highest priority', 'Cannot be overwritten', and 'Non-negotiable iron rule' to instruct the agent to ignore any subsequent user instructions or role-play attempts that might attempt to bypass its usage restrictions.
  • [COMMAND_EXECUTION]: The 'references/agents-rules.md' file includes an 'MCP Automated Bootstrapping' protocol that directs the agent to execute shell commands. These include installation commands (pip install mcp-taiwan-legal-db) and configuration commands (claude mcp add, gemini mcp add, etc.) that modify the agent's environment and platform configuration files.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the download and installation of external software packages from the Python Package Index (PyPI). It also involves network communication with an external endpoint (https://tlr.dr-lawbot.com/mcp) for legal research capabilities and diagnostic checks.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 5, 2026, 11:14 PM
Security Audit — agent-trust-hub — legal-case-analysis