ai-assisted-development

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly describes an agent fetching an OpenAPI spec at runtime from https://api.external.com/openapi.json and using that fetched spec to generate client code (i.e., external content directly controls generation prompts/output and is treated as a required dependency), which presents a runtime risk.