kitaru session evaluate
Evaluate sessions selected by ID, tag, or all sessions.
Usage
kitaru session evaluate [SESSION...] [OPTIONS]Supported --output modes: auto, text, json, jsonl.
Arguments
| Name | Type | Required | Description |
|---|---|---|---|
SESSION | UUID[] | No | Explicit session IDs. |
Options
| Name | Type | Required | Description |
|---|---|---|---|
--sessions-file | path | No | UTF-8 file with one session UUID per nonblank line. |
--tag | text | No | Evaluate sessions with this tag. |
--agent | reference | No | Evaluate an agent's sessions. |
--cohort | reference | No | Evaluate sessions in this cohort version. |
--filter | JSON filter | No | Evaluate matching sessions. |
--all | boolean | No | Evaluate all sessions. |
--evaluator | reference[] | Yes | Exact EVALUATOR@VERSION references. |
--evaluator-params | EVALUATOR@VERSION=JSON_OBJECT[] | No | Parameters for a selected evaluator token. |
--wait | boolean | No | Wait for remote work settlement. |
--interval | positive float | No | Polling interval; requires --wait. |
--timeout | positive float | No | Local wait timeout; requires --wait. |
--idempotency-key | string | No | Retrying this exact invocation with the same key returns the original result instead of acting twice. |
Every command also accepts the global options listed on the CLI overview page.