openapi-risk

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

• Third-party content exposure detected (high risk: 0.85). The skill calls OpenAPI endpoints (e.g., POST /<type> → GET /<type>/{id} → GET /<type>/{id}/download) to retrieve third-party credit/KYC reports and PDFs from risk.openapi.com, which are outsider-authored content (official registers/CRIF) that can be ingested as readable text into the agent’s LLM context.