kafka-shadowtraffic-java

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The required workflow runs the kafka-shadowtraffic skill for MCP-based topic/schema discovery, which can fetch outsider-authored free text from external sources (e.g., public web/MCP-provided content) at runtime and inject it into the agent’s LLM context during discovery/config generation.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly runs and depends on the remote ShadowTraffic Docker image (shadowtraffic/shadowtraffic:latest) at runtime (via the docker run lint command in Step 6 and as a TestContainers container), which pulls and executes remote code required for the skill to operate.