dev-java
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for standard Java and Spring development workflows. It mandates following repository-specific constraints and uses established tools like Maven and Gradle via their respective wrappers.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests and processes untrusted external data, such as Java source code and OpenAPI specifications.
- Ingestion points: The skill reads Java source files, Maven/Gradle manifests, and OpenAPI YAML files during the implementation and migration workflows (SKILL.md, references/protocol-contracts.md).
- Boundary markers: While it does not define specific LLM-level delimiters, it requires formal acceptance of contracts (OpenAPI) and uses repository-native seams for validation.
- Capability inventory: The skill is authorized to edit source files and execute build and test commands through Maven or Gradle (SKILL.md, references/java-engineering.md).
- Sanitization: The instructions emphasize standard application-level validation (e.g., Spring validation) and suggest normalizing generated outputs to detect unexpected drift, which helps mitigate accidental injection effects.
- [REMOTE_CODE_EXECUTION]: All referenced external URLs point to well-known technology organizations and official documentation repositories. These are used for reference and guidance, not for direct script execution via the command line.
- Evidence: References include official documentation for Spring, OWASP, and fixed snapshots of common reference projects like RuoYi-Vue and Spring Petclinic (references/java-engineering.md).
Audit Metadata