saas-msa-review
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted third-party contracts. * Evidence Chain: (1) Ingestion point: SaaS or subscription agreement text; (2) Boundary markers: Absent; (3) Capability inventory: Filesystem writes to local matter directories and invocation of research tools; (4) Sanitization: None identified.
- [DATA_EXFILTRATION]: The skill accesses and writes sensitive information, including legal playbooks and matter-specific files, to directories in the user's home folder (~/.claude/plugins/config/claude-for-legal/).
- [COMMAND_EXECUTION]: The instructions direct the agent to execute internal platform commands such as /commercial-legal:matter-workspace and /commercial-legal:cold-start-interview to manage state and context.
- [EXTERNAL_DOWNLOADS]: The skill is designed to interact with external legal research services (e.g., Westlaw) to verify statutory overrides and jurisdictional rules.
Audit Metadata