chinese-legal-law-student

Warn

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill instructs the agent to read from ~/.claude/plugins/config/..., which is a sensitive directory used by the Claude CLI to store user-specific configuration and potentially credentials. Accessing files outside the skill's workspace or in hidden tool configuration directories is a high-risk behavior for data exposure.
  • [PROMPT_INJECTION]: The skill utilizes a Codex adapter pattern that ingests instructions from external files such as law-student/CLAUDE.md. Ingestion points: law-student/CLAUDE.md, law-student/skills/*/SKILL.md, and ~/.claude/plugins/config/. Boundary markers: Absent. Capability inventory: Reading local files, web verification, and document rendering. Sanitization: Absent. This architecture makes the agent vulnerable to indirect prompt injection if the referenced files are sourced from an untrusted repository or modified by a malicious actor.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 17, 2026, 07:43 AM
Security Audit — agent-trust-hub — chinese-legal-law-student