---
title: "Task Status | AgentRQ Glossary"
url: https://agentrq.com/glossary/task-status
description: "Task status is the current lifecycle state of an AgentRQ task — pending, ongoing, completed, or failed — used to track agent progress and coordinate human oversight."
---

[Back to Glossary](https://agentrq.com/glossary/)

# Task Status

**Task status** is the current lifecycle state of an AgentRQ [task](https://agentrq.com/glossary/task). Status values let humans and agents track progress, identify blockers, and understand at a glance what work is active, waiting, or done.

## Status Values

| Status | Meaning | Agent Behavior |
| --- | --- | --- |
| `pending` | Task created, work not started | Agent may not have picked it up yet |
| `ongoing` | Agent is actively working | Real-time updates expected |
| `completed` | Work finished successfully | Agent has signaled done |
| `failed` | Unrecoverable error occurred | Human intervention likely needed |

## Status Transitions

A task typically moves through status in this order:

text

```text
task create (pending)
      ↓
task updateStatus (ongoing)
      ↓
task updateStatus (completed)  — or —  task updateStatus (failed)
```

A status change is one call: `task {"action": "updateStatus", "payload": {"taskId": "…", "status": "ongoing"}}`.

The agent is responsible for calling `task` with the `updateStatus` action at appropriate points in its work.

## Best Practices for Status Management

Following the AgentRQ MCP server instructions, agents should:

1. **Set `ongoing` immediately** when starting work on a received task
2. **Set `completed`** after sending a final summary via `task` reply
3. **Set `failed`** if an unrecoverable error prevents completion
4. **Never leave tasks in `pending`** indefinitely without action

## Reading Status on the Task Board

The [task board](https://agentrq.com/glossary/task-board) uses status to organize and prioritize what you see:

- → **Ongoing** tasks appear prominently — these need your attention
- → **Pending** tasks are queued — the agent will address them in order
- → **Completed** tasks form your audit trail
- → **Failed** tasks surface as alerts requiring human review

## Related Terms

- → [Task](https://agentrq.com/glossary/task)
- → [Task Board](https://agentrq.com/glossary/task-board)
- → [MCP Tool](https://agentrq.com/glossary/mcp-tool)
- → [Human-in-the-Loop](https://agentrq.com/glossary/human-in-the-loop)
- → [Notification](https://agentrq.com/glossary/notification)