522-frameworks-micronaut-testing-integration-tests
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user requests to generate or refactor integration tests, creating a surface for indirect instructions. 1. Ingestion points: User prompts requesting new tests or infrastructure wiring (SKILL.md). 2. Boundary markers: Absent; the skill lacks instructions to delimit or ignore malicious directives within user input. 3. Capability inventory: The skill requires shell command execution via Maven and has file system write access for Java code. 4. Sanitization: Absent; no validation or filtering of user-supplied content is defined.
- [COMMAND_EXECUTION]: The skill workflow requires the agent to execute shell commands, specifically
./mvnw compileandmvn clean verify, to ensure project state and validate changes.
Audit Metadata