api-to-bruno

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the LLM context when the skill resolves a “direct spec URL” or “Git URL” and then reads OpenAPI/Swagger/WSDL or route-defining files from that external source (public web content / third-party repo) to confirm markers and inventory routes.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill will clone or fetch user-provided Git or OpenAPI/WSDL URLs at runtime (e.g., a user-supplied Git URL like https://github.com/org/repo.git or a direct spec URL such as https://example.com/openapi.yaml) and then reads that fetched content to drive generation of Bruno files, meaning external content directly controls the agent's output/context.