The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents methods for reading external, untrusted data from Telegram chats and channels, creating a surface for potential indirect prompt injection attacks.
Ingestion points: The skill uses iter_messages, iter_dialogs, and iter_participants to fetch messages and metadata from the Telegram platform (references/entities-and-history.md).
Boundary markers: No delimiters or explicit instructions to treat message content as untrusted are provided in the implementation examples.
Capability inventory: The skill describes capabilities for sending messages (send_message) and downloading files (download_media), which could be triggered by processed data (SKILL.md).
Sanitization: The documentation does not demonstrate sanitization or filtering of incoming message text before processing.

telegram-userbots