The Agent Skills Directory

[PROMPT_INJECTION]: The skill describes functionality for reading and processing Telegram message history via iter_messages and iter_dialogs. This creates a surface for indirect prompt injection, where malicious instructions embedded in external messages could influence the agent's behavior.
Ingestion points: Untrusted message data enters the context from the Telegram network as described in references/entities-and-history.md.
Boundary markers: The documentation does not provide specific delimiters or instructions to ignore embedded commands in the message history.
Capability inventory: The skill possesses network communication capabilities via SKILL.md and file system access for session management as described in references/sessions-and-auth.md.
Sanitization: There are no instructions for sanitizing or validating the content of messages before processing.
[EXTERNAL_DOWNLOADS]: The skill relies on the telethon Python library for MTProto communication. It recommends strict version pinning (e.g., telethon==1.44.0) to ensure stability across protocol updates.
[CREDENTIALS_UNSAFE]: The documentation discusses the management of Telegram session strings and API credentials. While it advocates for secure storage in environment variables and secret stores, these artifacts represent full account access and are primary targets for exposure if handled incorrectly.

telegram-userbots