read_output
Read-only
Retrieve new output from a running command without sending input. Use after a partial status to continue monitoring long-running processes.
Instructions
Read new output from a running command without sending any input.
Use this after execute returns status "partial" when the command needs no
interaction (e.g. long-running build, training loop, find).
Returns the same dict as execute:
status="partial": output + command_id; poll again for more.
status="completed": output + exit_code. The command_id is now spent —
it is closed and removed, so a later read_output on
it raises "Invalid command_id".
Each call returns only the output produced since the previous call.
Args:
- pause_timeout: seconds of OUTPUT SILENCE before returning (default 9.0).
This is the dial that controls how long a silent-poll call waits.
Raise it (e.g. 30, 60) when polling a very quiet job; raising
total_timeout instead does nothing while the process stays silent.
- total_timeout: hard cap on call duration (default 20.0). Only binds
while output is actively streaming.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| command_id | Yes | The active command_id returned in a partial status response | |
| pause_timeout | No | Seconds of output silence to wait before returning (default is 9.0) | |
| total_timeout | No | Hard cap in seconds on the maximum duration of this call (default is 20.0) |