x-twitter-scraper

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

• Third-party content exposure detected (medium risk: 0.65). At runtime, the skill can ingest X-authored free text from webhook/event payloads (e.g., POST /webhooks deliveries or GET /events), which are outsider-authored content and become LLM-readable context unless isolated.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly directs agents to fetch live authoritative sources at runtime—notably the MCP endpoint (https://xquik.com/mcp) and the machine-readable OpenAPI spec (https://xquik.com/openapi.json)—and those runtime fetches return endpoint metadata/spec that directly shape tool selection and agent prompts, so they are runtime external dependencies that can control agent instructions.