Thanks — I’m responding to your question about whether an agent task’s

complete()

function is visible to the LLM and can be called as a tool. Short answer: No —

complete()

is not visible or callable by the LLM as a tool. Why: •

complete()

is an internal method the agent task uses to finalize its execution and return a result to the controlling agent. It isn’t exposed to the LLM runtime. • If you want the LLM to call functionality, you must define explicit tools using

@function_tool

(Python) or

llm.tool()

(Node.js). Those tool functions can perform work and — if needed — call

complete()

internally, but the LLM only sees and invokes the tool function itself, not

complete()

. If it helps, I can: • Show a short example of how to wrap task-finalization in a tool function (Python or Node) so the LLM can call it. • Review a snippet of your current task/tool code and suggest how to expose the behavior you want as a callable tool. Which would you prefer? Sources: Tasks and task groups | LiveKit Documentation | Tool definition and use | LiveKit Documentation