Skip to content

Using the Task Queue in Develop Mode

The Task Queue is where your plain-language requests become real code changes. This guide covers how to write effective task requests, manage queue order, and understand execution outcomes.

Applies to: Develop Mode — Prompt Engineer and Task Queue agents


The task lifecycle

Your request → Prompt Engineer → Task Queue → Coding Agent → Commit
StateMeaning
PendingTask created, not yet scheduled
QueuedWaiting for an available coding agent
RunningAgent is actively reading and modifying your codebase
CompletedCode changes committed and pushed to your branch
FailedAn error stopped execution — view details and retry

Tasks within the same project run sequentially on the same branch to avoid merge conflicts.


Writing effective task requests

The Prompt Engineer refines your request, but the better your input, the better the output.

Be specific about the target

Instead of:

“Improve the database queries”

Write:

“Fix the N+1 query in OrderListView — use select_related('customer', 'items__product') to reduce the query count on the order listing endpoint”

Describe the outcome, not the implementation

Instead of:

“Add a Redis cache”

Write:

“Cache the product catalogue endpoint (GET /api/products) with a 5-minute TTL to reduce database load. Invalidate the cache when any product is updated or deleted.”

Specify scope boundaries

If you want the agent to stay within a specific file or module:

“Only modify files in billing/. Don’t change any API endpoint contracts.”

Include testing requirements

“Write pytest tests for the new caching logic. Use pytest-mock to mock Redis calls. Coverage should reach 90% on the modified functions.”


Reviewing the optimised prompt

After you submit your request, the Prompt Engineer shows you the optimised version before it’s queued. Always read it:

  • Does it target the right files?
  • Are the scope boundaries correct?
  • Does it include the testing instruction you wanted?

Edit the prompt inline if anything is off, then click Add to Task Queue.


Managing the queue

Viewing the queue

The Task Queue panel shows all tasks grouped by status. For each task you can see:

  • Title and brief description
  • Current status and elapsed time (for running tasks)
  • Commit hash and message (for completed tasks)

Cancelling a task

Tasks in Pending or Queued state can be cancelled. Click the ✗ button next to the task. Tasks already Running cannot be cancelled mid-execution.

Re-ordering the queue

Drag tasks in the Pending section to reorder them. The coding agent picks the next task from the top of the Pending list when it finishes the current one.


Interpreting results

Completed tasks

For each completed task:

  • Commit hash — links to the commit on your Git host
  • Files modified — list of changed files
  • Summary — a human-readable description of what was changed

Review the diff on your Git host before merging to a protected branch.

Failed tasks

When a task fails:

  1. Click View error to see the full error message and stack trace.
  2. Common causes:
    • Compilation error — the generated code has a syntax or type error
    • Test failure — the agent wrote code that doesn’t pass its own tests
    • File not found — the target file was moved or renamed since the task was created
    • Git push rejected — a conflicting push happened on the same branch
  3. Retry the task after reading the error — many failures are automatically resolved on retry.
  4. If the retry fails again, update the task description via the Prompt Engineer with additional context about the error.

Task capabilities

The coding agent can handle a broad range of tasks:

CategoryExamples
RefactoringExtract function, rename symbol across codebase, restructure module
Bug fixesFix a specific error given a stack trace
Feature implementationAdd an endpoint, implement a service, integrate an SDK
Test generationUnit tests, integration tests, edge case coverage
DocumentationDocstrings, README updates, OpenAPI annotations
Dependency updatesUpgrade a package and fix breaking changes
PerformanceOptimise a query, add caching, reduce memory usage

Next steps