502-frameworks-micronaut-rest

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires the agent to run local shell commands (./mvnw compile, mvn clean verify) to validate project state and verify REST API improvements. These operations are standard and expected within a developer-oriented skill for Java/Micronaut environments.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it ingests untrusted project data and has shell execution capabilities.
  • Ingestion points: The skill analyzes project source code (controllers, DTOs) and OpenAPI specification files.
  • Boundary markers: There are no specific delimiters or markers used to isolate untrusted code content during processing.
  • Capability inventory: The skill utilizes shell command execution via the Maven wrapper (./mvnw).
  • Sanitization: No explicit sanitization or filtering of the ingested source code is documented before the analysis phase.
  • [SAFE]: The skill demonstrates a strong security posture by instructing developers to implement centralized error handling that avoids leaking stack traces to clients in production.
  • [SAFE]: The instructions encourage the use of Jakarta Bean Validation and security annotations (@Secured) to maintain robust security boundaries at the API layer.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 06:06 PM