integration-test-grill
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill acts as an instructional alias for testing strategies.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied product flow descriptions as part of its core functionality.
- Ingestion points: Product flow descriptions processed via the default_prompt in agents/openai.yaml and instructions in SKILL.md.
- Boundary markers: No delimiters or explicit instructions to ignore embedded content are defined.
- Capability inventory: No dangerous capabilities such as subprocess execution, network operations, or file system writes are present in the skill files.
- Sanitization: No input sanitization or validation is implemented.
Audit Metadata