xapi

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to call public third‑party endpoints (e.g., twitter.* calls like "npx @xapi-to/xapi call twitter.tweet_detail", web.search/realtime, news.search.latest, and lists third‑party services such as Reddit and Serper), meaning the agent will fetch and read untrusted, user‑generated web and social content that can influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill requires running the remote package via "npx @xapi-to/xapi", which fetches and executes code from the npm registry at runtime (remote code execution) and is a required dependency for the skill.

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

• Direct money access detected (high risk: 1.00). The skill exposes explicit payment-related actions: notably the "topup" command with a Stripe method (npx @xapi-to/xapi topup --method stripe --amount 10), which is a payment gateway integration allowing money to be moved/charged. That qualifies as Direct Financial Execution. (While crypto endpoints shown are read-only price/metadata calls, the presence of a Stripe top-up flow is sufficient to flag financial execution capability.)