integration-e2e-testing
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to process external, potentially untrusted sources including Design Docs, Product Requirement Documents (PRDs), and UI Specifications to calculate ROI and generate test skeletons.
- Ingestion points:
SKILL.md(ROI Calculation section) andreferences/e2e-design.md(Candidate Sources section) reference processing PRDs, Design Docs, and UI Specs. - Boundary markers: The skill does not specify delimiters or instructions to help the agent distinguish its core logic from the content of processed external documents.
- Capability inventory: The skill is designed to guide agents that may have file-writing or code-generation capabilities to produce the 'Test Skeletons' defined in the specification.
- Sanitization: No sanitization, validation, or escaping of input document content is described, allowing potentially malicious instructions in a Design Doc to influence the agent's generation of test code.
Audit Metadata