Task Workspace
Documentation and guides for OutcomeDev.
This document provides a comprehensive overview of the task workspace UI, framed as a user journey from signing in to completing a task.
1. Sign In & Connect
Your journey begins on the home page. Before you can start a task, you need to sign in and connect your GitHub account.
- Sign In: Use the sign-in buttons to authenticate with your preferred provider (e.g., GitHub, Vercel).
- Connect GitHub: If you haven't already, you'll be prompted to connect your GitHub account. This allows the application to access your repositories.
2. Select or Create a Repository
Once you're signed in, you can select a repository to work on.
- Owner & Repo Dropdowns: Use the dropdowns in the header to select the owner (your personal account or an organization) and the repository.
- Create New Repository: If you want to start a new project, you can use the "New" button to create a new repository on GitHub directly from the application.
3. Define Your Outcome
This is where you tell the application what you want to achieve.
- Suggested Outcomes: The application will analyze your repository and suggest a few possible next steps. You can choose one of these to get started quickly.
- Write Your Own: If you have a specific task in mind, you can click "or write your own outcome" to open a prompt editor.
4. The Task Workspace
Once you start a task, you'll be taken to the task workspace. This is your mission control for the task.
Header
The header provides a high-level overview of the task.
- Task Status: Shows the current status of the task (e.g.,
processing,completed). - Branch Name: The Git branch created for this task.
- Pull Request (PR) Status: Once a PR is created, you'll see its status (
open,closed,merged), a link to the PR on GitHub, and the status of any CI/CD checks. - Sandbox Status: The status of the isolated environment where the code is being executed.
- Agent & Model: The AI agent and model working on the task.
- Task Actions: A dropdown menu with key actions:
- Create PR: Manually create a pull request with the current changes.
- Retry Task: If a task fails, you can retry it. You can also modify the task parameters (e.g., change the model, install dependencies) before retrying.
- Delete Task: Permanently delete the task.
Panes
The workspace is divided into several resizable panes.
- Files Pane: Navigate your project's file tree.
- Files View: Browse all files in the repository.
- Changes View: See only the files modified by the current task.
- Code Pane: Your primary code editor. It features a tabbed interface, syntax highlighting, and a diff viewer to see changes.
- Preview Pane: A live, interactive preview of your application. It reloads automatically as the agent makes changes.
- Chat Pane: Communicate directly with the AI agent. You can provide feedback, ask questions, and guide the task.
Bottom Bar
- Logs Pane: Streams the raw output from the task's execution environment, including commands run by the agent and any errors.
5. Review & Merge
Once the task is complete, you can review the changes and merge them into your main branch.
- Create Pull Request: If the agent hasn't already, you can create a pull request from the "Task Actions" menu.
- Review Checks: In the header, you can monitor the status of your CI/CD checks. If any checks fail, you can use the Chat pane to ask the agent to fix the issues.
- Merge Options: Once you're ready to merge, you'll have several options:
- Squash and Merge: Combine all of the PR's commits into a single commit on the main branch.
- Create a Merge Commit: Add all of the PR's commits to the main branch in a single merge commit.
- Rebase and Merge: Rebase the PR's commits onto the main branch and then merge them.
- Conflict Resolution: If there are merge conflicts, a dialog will appear. You can choose to have an AI agent attempt to resolve the conflicts for you.