quality-engineering-test-healing
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill establishes a safe operating environment by providing a clear failure taxonomy and a list of forbidden repairs. It prevents security-sensitive or quality-degrading actions such as weakening assertions, inflation of timeouts, or touching production code.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze untrusted artifacts such as test reports, DOM diffs, and traces which could theoretically contain malicious content designed to influence agent behavior. However, the risk is mitigated by the skill's rigid framework of allowed actions and verification requirements.
- Ingestion points: Test artifacts located in
test-results/**andplaywright-report/**. - Boundary markers: None explicitly defined in the provided instructions.
- Capability inventory: No execution capabilities (shell access, network operations, or file-write) are requested or defined in this skill set.
- Sanitization: No specific sanitization or filtering of test output is mentioned.
Audit Metadata