skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/payment-application-processor/Gen Agent Trust Hub
payment-application-processor
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external project data provided by the user in formats such as CSV, Excel, and JSON to generate payment reports. This ingestion of untrusted data presents a surface for indirect prompt injection if malicious instructions are embedded within the construction data.
- Ingestion points: The
instructions.mdfile specifies that the agent should accept data via file paths or direct input in common data formats. - Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings for the agent to use when processing this data.
- Capability inventory: The skill requests
filesystempermissions inclaw.json, allowing the agent to read and write project data. - Sanitization: While
instructions.mdincludes a general constraint to "Validate inputs before processing," it lacks technical specifications for sanitizing or escaping content from external files.
Audit Metadata