Skip to main content

Run dynamic agent tasks

Use the Dynamic Agent Tasks panel to configure, run, and duplicate AI-powered tasks with custom inputs and file or report attachments.

Audience

This guide is designed for project manager.

Steps​

Step 1: Open the Dynamic Agent Tasks panel​

Navigate to the relevant agent view in your project and open the Dynamic Agent Tasks panel to see the list of available tasks for that agent.

Open the Dynamic Agent Tasks panel

Step 2: Choose a task to run​

Review the task list, including icons, titles, descriptions, and IU cost badges, then select the task you want to run from the panel.

Choose a task to run

Step 3: Review task details and IU cost​

Open the selected task to see its detailed description, required inputs, and IUCostBadge, and check for any Early Access badge if applicable.

Review task details and IU cost

Step 4: Fill in text and option fields​

Complete the required and optional input fields shown by the DynamicInputField components, entering text, selecting from dropdowns, or using searchable selects as needed.

Fill in text and option fields

Step 5: Attach reports from your project​

For report picker fields, search and select one or more existing reports from your project so the agent can use them as context when running the task.

Attach reports from your project

Step 6: Upload supporting files​

For file upload fields, drag and drop files into the upload area or use the upload button to attach single or multiple files, then remove or clear files if you change your mind.

Upload supporting files

Step 7: Search and filter dynamic options​

When a field supports dynamic option search, start typing to filter the available options, and select the most relevant one from the SearchableSelect dropdown.

Search and filter dynamic options

Step 8: Apply default values and review inputs​

Check any prefilled values that have been set using default expressions, adjust them if required, and confirm that all mandatory fields are complete.

Apply default values and review inputs

Step 9: Start the task​

Click the button to start the task; the panel validates your inputs and triggers the task execution, while onTaskStart passes the task and project details to the backend.

Start the task

Step 10: Monitor task progress​

Watch the panel as the task runs, with loading indicators such as Loader2 and status icons showing when the task succeeds or fails.

Monitor task progress

Step 11: View task results and reports​

Once complete, review the TaskExecutionResult, including any generated report summaries, links, or structured outputs made available by the agent.

View task results and reports

Step 12: Duplicate a successful task​

Use the duplicate action (Copy icon) to clone a previous task run, automatically reusing or adjusting its inputs so you can run a similar task more quickly.

Duplicate a successful task

Step 13: Build a new custom task​

Open the AgentTaskCreatorDialog from the panel to define a new custom task for the current agent, including its name, description, and input fields.

Build a new custom task

Step 14: Save your custom task​

Finalise the configuration in the task builder, save your new task, and let the panel refresh so the new task appears in the available task list.

Save your custom task

Step 15: Close the task creator​

When you are finished editing or creating tasks, close the AgentTaskCreatorDialog using the close control, which triggers handleTaskCreatorClose and returns you to the main panel.

Close the task creator

Step 16: Handle authentication or access issues​

If you are signed out or lack access, respond to any notifications raised by notifySignedOut or error toasts, then sign in again and retry your task.

Handle authentication or access issues

Last updated: 2026-06-29