test-design-techniques
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data such as requirements, code, and API specifications to generate test cases, creating a potential surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the context through inputs defined in
schemas/output.json(requirements, code, api-spec, user-story). - Boundary markers: No specific delimiters or "ignore instructions" markers are present to separate untrusted data from instructions.
- Capability inventory: The skill coordinates multiple agents (qe-test-generator, qe-coverage-analyzer, qe-quality-analyzer) and generates test artifacts in various formats (JSON, YAML, Markdown, HTML, CSV).
- Sanitization: No explicit validation or sanitization of input data is defined in the skill logic to prevent instruction leakage from untrusted input.
Audit Metadata