custom-telegram-bot
Pass
Audited by Gen Agent Trust Hub on Jun 24, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection through its message polling functionality.
- Ingestion points: Incoming messages and media are retrieved via the get_updates action defined in SKILL.md and schema.md.
- Boundary markers: The instructions do not specify any delimiters or warnings to the agent regarding the untrusted nature of external message content.
- Capability inventory: The skill provides broad capabilities to interact with the environment, including sending messages (send_message), sharing files (send_photo, send_document), and managing files via the AgentPMT File Manager.
- Sanitization: There is no evidence of data sanitization or validation performed on incoming Telegram messages before they are processed by the agent.
- [EXTERNAL_DOWNLOADS]: The skill documentation specifies the use of the platform's package manager to install supporting skills from the vendor's official repository.
Audit Metadata