dpa-review
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of DPA documents (PDFs, links, or text provided by third parties).
- Ingestion points: Reads external files (DPA documents) provided via user arguments or local file paths.
- Boundary markers: The instructions include explicit guidelines for source attribution and verification (e.g., labels like
[已确定],[需验证]) and prohibit the model from silently supplementing legal content without retrieval. - Capability inventory: The skill possesses the ability to read and write to local files in the
~/.claude/directory and perform external information retrieval for legal research. - Sanitization: The skill relies on process-based labeling and requires human (lawyer) oversight rather than performing automated input sanitization.
Audit Metadata