telegram
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates the creation of Telegram bots that ingest untrusted user input, presenting a surface for Indirect Prompt Injection attacks.
- Ingestion points: User messages are ingested through the Telegram Bot API and processed in
SKILL.md(Automacao Com Ia section),assets/boilerplate/python/bot.py, andassets/boilerplate/nodejs/src/handlers.ts. - Boundary markers: No delimiters or defensive instructions (e.g., 'ignore embedded instructions') are used when interpolating user text into AI prompts in the provided examples.
- Capability inventory: The provided code and documentation enable the bot to send messages, photos, and documents, as well as manage group members and chat settings via the Telegram Bot API.
- Sanitization: The example code lacks input validation or sanitization before passing user content to external AI services (e.g., the Anthropic client snippet).
Audit Metadata