contract-clause-risk-scanner
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (contract clauses), creating a surface for indirect prompt injection if the content is later used in LLM prompts.
- Ingestion points: run.py reads JSON/CSV data from a path provided via the --input argument.
- Boundary markers: The current implementation does not include markers to delimit untrusted content.
- Capability inventory: The script performs file read/write operations (Path.read_text and Path.write_text) but does not currently execute the content or pass it to an LLM.
- Sanitization: No sanitization or validation of the input text is performed before processing.
Audit Metadata