legal-article-retrieval

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a template for legal research and does not contain any executable code or malicious instructions.
  • [PROMPT_INJECTION]: No evidence of prompt injection or safety bypass attempts was found. The skill uses directive language ('MUST', 'Strictly Prohibited') to ensure the agent adheres to factual legal sources and avoids hallucination.
  • [CREDENTIALS_UNSAFE]: The README.md file contains a configuration template for an MCP server that uses standard placeholders like 'YOUR_TOKEN' and '{SERVICE_ID}'. It explicitly warns users not to commit actual credentials to repositories.
  • [EXTERNAL_DOWNLOADS]: The skill mentions connecting to external legal databases via MCP. The primary example points to 'pkulaw.com', which is a well-known and established legal research service in China. This is documented neutrally as a configuration step for the user.
  • [DATA_EXFILTRATION]: There are no commands or logic designed to harvest sensitive user files (such as SSH keys or environment variables) or send data to unauthorized external endpoints.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:56 AM
Security Audit — agent-trust-hub — legal-article-retrieval