x-twitter-scraper

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and processes untrusted, user-generated content from public third-party sources—e.g., tweets and articles via GET /x/tweets/{id} and search endpoints, and external news/social sources via the Radar endpoints (references/radar, SKILL.md Retrieval Sources and MCP tooling)—and the agent is instructed to read and analyze that content as part of normal workflows (search, compose, extractions, monitors), so that content could indirectly influence tool selection or actions.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly includes billing endpoints and actions (e.g., "POST /credits/quick-topup", "Billing Guide", quick top-up, credits/subscriptions, and MPP endpoints) and instructs the agent how to call them. Those are specific, non-generic payment-related operations that can move money/credits (even though they require explicit approval). This constitutes direct financial execution capability.