Yeah but this seems like a bolt-on and not something they train their model to understand at the token level like how they do tool calls. Maybe Anthropic has a

Yeah but this seems like a bolt-on and not something they train their model to understand at the token level like how they do tool calls. Maybe Anthropic has a token-level skills support (e.g. <SKILL_START>skill prompt<SKILL_END>).