The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user data from Telegram messages and forwards it to the agent, creating a surface for indirect prompt injection.
Ingestion points: Telegram updates arrive via telegram-bot::webhook in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are mentioned in the message processing logic.
Capability inventory: The skill triggers agent turns via harness::send and manages tool approvals through approval-gate as documented in SKILL.md.
Sanitization: There is no evidence of sanitization or filtering applied to incoming user text before prompt interpolation.

telegram-bot