Missions
Missions are the core workflow in CoFounderBot. A mission is a high-level goal that gets broken into steps and assigned to specialized agents.
Creating a Mission
Go to Dashboard → Missions → New Mission and describe what you want to build. Be specific — the more context you provide, the better the output.
Good Mission Descriptions
``` Build a Next.js landing page for a SaaS product called "DataFlow". Include: hero section with CTA, features grid (6 features), pricing table (3 tiers), FAQ accordion, and footer. Use Tailwind CSS, make it responsive, dark theme. ```
Bad Mission Descriptions
``` Make a website ```
Tip: Include technology preferences, design requirements, and specific features. Agents work best with clear instructions.
Execution Modes
Plan & Execute (Recommended)
The Director agent first creates a structured plan with steps like:
- Research → Architecture → Coding → Testing → Consolidation → Quality Check
Each step is assigned to the best agent for the job and executed in order, respecting dependencies.
Execute Only
Skips the planning phase. The task goes directly to the most appropriate agent based on its content. Best for simple, single-agent tasks.
Mission Lifecycle
| Status | Meaning |
|---|---|
| `draft` | Created but not started |
| `running` | Agents are actively working |
| `blocked` | A step needs human input |
| `completed` | All steps finished successfully |
| `failed` | A step failed after retries |
Step Details
Each step shows:
- Agent assigned (e.g., Forge for coding)
- Status (pending → running → completed/failed)
- Output with Preview, Files, and Tools tabs
- Rating buttons for feedback
GitHub Auto-Commit
If you configure a GitHub token in Settings, the Coder agent automatically commits code to a branch named `mission-{id}`. You can then create a PR directly from the mission page.
Tips
- Use Plan & Execute for complex projects (multiple files, tests, architecture)
- Use Execute Only for quick tasks (fix a bug, write a function, explain code)
- Rate outputs to help agents improve over time via the self-improvement system
- Check the Team page to see agent performance trends