subpoena-triage
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided legal documents (Step 1), creating a potential surface for instructions hidden within the data to influence agent behavior.
- Ingestion points: Step 1 reads subpoena and investigation notice files provided by the user.
- Boundary markers: There are no explicit instructions to delimit the document content or warn the agent to ignore embedded commands.
- Capability inventory: The skill writes triage reports to a local plugin configuration directory (~/.claude/plugins/config/claude-for-legal-zh/) and invokes other skills such as /legal-hold, /matter-intake, and /matter-briefing.
- Sanitization: The workflow does not include steps to sanitize or filter external content before processing it or using it to trigger subsequent actions.- [COMMAND_EXECUTION]: The skill automatically routes tasks to other agent capabilities (e.g., /legal-hold, /matter-intake) based on its analysis of the input document, which could lead to unintended tool execution if the input document contains malicious content.
Audit Metadata