test-implement
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of instructional content for standard frontend testing practices. It follows the AAA (Arrange-Act-Assert) pattern and promotes test independence and deterministic execution.
- [SAFE]: The documentation includes explicit security best practices, such as ensuring test data is minimal and free of real sensitive values in the
references/frontend.mdfile. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze a repository's existing configuration, scripts, and fixtures to execute tests and services. This constitutes a vulnerability surface common to coding tools where instructions embedded in the repository could influence agent behavior.
- Ingestion points: Repository browser-test configurations, package scripts, fixtures, and neighboring tests identified in
references/e2e.mdandreferences/frontend.md. - Boundary markers: None identified.
- Capability inventory: The skill instructs the agent to identify and use service start commands and health-check commands, which implies command execution capabilities.
- Sanitization: The skill explicitly instructs to keep test data "free of real sensitive values."
Audit Metadata