insurance-agent-gap-advisor
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (financial, health, and family details) which is used to perform calculations and drive calls to external tools.
- Ingestion points: User input described in the
SKILL.mdprerequisites and Step 1 ofassets/workflow.md. - Boundary markers: The skill includes logic to confirm information before proceeding, but lacks explicit delimiters for user-provided strings.
- Capability inventory: Uses
customer-system.get_customer_policiesandproduct-system.search_productsto fetch data based on user identity and profile. - Sanitization: Employs
scripts/validate_input.pyto check the validity of age, income, and family structure inputs. - [COMMAND_EXECUTION]: The skill executes a local Python script (
scripts/validate_input.py) to perform data validation during the workflow. This execution is intended for internal logic and does not expose a shell for arbitrary command injection.
Audit Metadata