legal-nda-triage
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided legal text, which introduces a vulnerability surface for indirect prompt injection attacks where malicious instructions could be embedded in the document text.
- Ingestion points: Untrusted data enters the agent context via user-supplied text when requesting a review of an NDA (e.g., '투자사에서 보내온 NDA 검토해줘... [NDA 내용]').
- Boundary markers: The skill does not define specific delimiters or 'ignore' instructions to separate user-provided document text from its own logic.
- Capability inventory: The skill primarily performs text analysis and generates structured reports; it lacks dangerous capabilities such as direct file system writes, network exfiltration, or shell command execution.
- Sanitization: There is no evidence of input sanitization, filtering, or validation for the document content processed by the skill.
Audit Metadata