china-legal-skills
Pass
Audited by Gen Agent Trust Hub on Jul 4, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs the
subprocessandchild_processmodules to execute local command-line utilities likepdftotextfor PDF-to-text conversion andosascriptto facilitate communication between internal skill components. These operations are restricted to processing local files provided by the user within the designated workspace. - [SAFE]: Network communication is limited to the legitimate legal database provider
pkulaw.com. This access is used to verify the current validity of statutes and regulations cited in legal documents. The skill correctly instructs users to manage API credentials through environment variables rather than hardcoding them. - [SAFE]: The skill implements a sophisticated 'Legal Formal Deliverable Hard Gate' system. This includes pre-export checks (
preflight_check.py) and post-export structural verification (health_check.py) to ensure that generated Word documents meet strict legal formatting standards and contain verified content. - [SAFE]: Analysis of the Python and JavaScript code revealed no evidence of obfuscation, credential exfiltration, or attempts to bypass agent safety guidelines.
Audit Metadata