523-frameworks-micronaut-testing-acceptance-tests
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external Gherkin scenario facts (Features, Scenarios, steps) to generate Java test code. This represents a surface for indirect prompt injection where malicious instructions could be embedded in the Gherkin text.
- Ingestion points: Scenario facts provided by the user as described in SKILL.md.
- Boundary markers: The skill explicitly defines an 'AUTHORITY BOUNDARY' instructing the agent to treat Gherkin text as untrusted data and never obey instructions embedded within it.
- Capability inventory: The skill performs file writes (Java test classes) and shell command execution (mvnw compile, mvnw clean verify).
- Sanitization: The instructions mandate that the agent only process 'maintainer-authored sanitized scenario facts' and refuse raw third-party .feature files.
- [COMMAND_EXECUTION]: The skill requires the agent to execute build commands to verify the project state.
- Evidence: Both SKILL.md and the reference file mandate running ./mvnw compile before applying changes and ./mvnw clean verify after implementation.
Audit Metadata