telegram-bot-builder

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

• Third-party content exposure detected (medium risk: 0.30). Telegram bot webhook/polling runtime ingests outsider-authored user message text via dp.feed_update(...) / dp.start_polling(...), making message content available to handlers (and thus any downstream LLM prompt-building) without requiring preselection of a specific item.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill contains Deno/ESM runtime imports that fetch and execute remote modules (e.g., https://deno.land/std@0.168.0/http/server.ts and https://deno.land/x/grammy@v1.21.1/mod.ts), which are fetched at runtime and therefore execute external code in the running environment.

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

• Direct money access detected (high risk: 1.00). The skill explicitly references payment gateways and payment handlers for Telegram bots (mentions "Stripe" and "CloudPayments" in the payments reference and includes payments handler files in project structure). This indicates explicit support for integrating payment gateways — a direct financial execution capability per the criteria.