chinese-legal-litigation
Warn
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXPOSURE]: The skill instructs the agent to access and read files from
~/.claude/plugins/config/. This directory is a sensitive path as it typically contains configuration profiles, user preferences, and potentially authentication tokens or session data related to the Claude Code tool. - [INDIRECT_PROMPT_INJECTION]: The skill operates as an adapter that dynamically reads and follows instructions from local/external files (e.g.,
litigation-legal/CLAUDE.mdand variousSKILL.mdfiles in theskills/subdirectory). This creates an attack surface where malicious instructions embedded in those files could influence the agent's behavior. - Ingestion points: Workflow files located at
litigation-legal/CLAUDE.mdandlitigation-legal/skills/*/SKILL.md. - Boundary markers: Absent; the instructions do not specify any delimiters or warnings to ignore embedded instructions in the sourced files.
- Capability inventory: The agent is instructed to use Codex tools for local file access, web verification, document rendering, and generating user-visible output.
- Sanitization: Absent; there is no mention of sanitizing or validating the content of the sourced legal workflows before execution.
Audit Metadata