chinese-legal-corporate
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXPOSURE]: The skill instructs the agent to read and potentially create configuration profiles located in hidden directories within the user's home folder, specifically
~/.claude/plugins/config/...and~/.codex/legal-zh/corporate-legal/CLAUDE.md. While these are used for cross-platform compatibility, accessing files in the user's home directory outside of the active project workspace is a privacy and data exposure concern. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process natural-language requests and external legal domain content (M&A documents, contracts, and resolutions), which serves as a vulnerability surface for indirect instructions.
- Ingestion points: User-provided requests,
corporate-legal/CLAUDE.md, and referenced skill files incorporate-legal/skills. - Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the legal documents being processed.
- Capability inventory: The skill utilizes tools for local file access, web verification, and document rendering.
- Sanitization: There are no mentioned mechanisms for sanitizing or validating the content of the legal documents before they are processed by the agent.
Audit Metadata