status subcommand gives a quick read-out of where a task currently stands: which required files are present, what the overall statuses in evidence.json and verdict.json are, and which criteria are not yet passing.
Unlike validate, status always exits with code 0. Use validate when you need a strict pass/fail check; use status when you want a human- or agent-readable summary during work.
Usage
Flags
Task identifier to summarize.
Optional working directory inside the repo. Defaults to the current directory. The script walks up from this path to discover the git repository root.
JSON output
status prints a JSON object to stdout. All fields are always present.
| Field | Type | Description |
|---|---|---|
repo_root | string | Absolute path to the discovered repository root. |
task_id | string | The requested task identifier. |
task_dir | string | Absolute path to .agent/tasks/<TASK_ID>/. |
exists | boolean | Whether .agent/tasks/<TASK_ID>/ exists. |
required_files_present | object | Map of each required file path (relative to task_dir) to a boolean indicating whether it exists. |
evidence_overall_status | string or null | Value of overall_status from evidence.json, or "PARSE_ERROR: <message>" if the file cannot be parsed, or null if the file does not exist. |
verdict_overall_status | string or null | Value of overall_verdict from verdict.json, or "PARSE_ERROR: <message>" if the file cannot be parsed, or null if the file does not exist. |
non_pass_criteria | array | List of criteria from verdict.json whose status is FAIL or UNKNOWN. Each entry has id, status, and reason. Empty when all criteria pass or when verdict.json is absent or unparseable. |