workgraph
First: run wg quickstart
Always run wg quickstart in your terminal at the start of a session. It prints
the current workflow cheat sheet and tells you whether the service is already running.
bash
wg quickstart
Your role as a top-level agent
You are a coordinator. Your job is to define work and let the service dispatch it.
Start the service if it's not running
bash
wg service start --max-agents 5
Define tasks with dependencies
bash
wg add "Design the API" -d "Description of what to do" wg add "Implement backend" --blocked-by design-the-api wg add "Write tests" --blocked-by implement-backend
Monitor progress
bash
wg list # All tasks with status wg agents # Who's working on what wg service status # Service health
What you do NOT do as coordinator
- •Don't
wg claim— the service claims tasks automatically - •Don't
wg spawn— the service spawns agents automatically - •Don't work on tasks yourself — spawned agents do the work
If wg done fails, the task may require verification — use wg submit.
If you ARE a spawned agent working on a task
You were spawned by the service to work on a specific task. Your workflow:
bash
wg show <task-id> # Understand what to do wg context <task-id> # See inputs from dependencies wg log <task-id> "msg" # Log progress as you work wg done <task-id> # Mark complete when finished
If you discover new work while working:
bash
wg add "New task" --blocked-by <current-task>
Manual mode (no service running)
Only use this if you're working alone without the service:
bash
wg ready # See available tasks wg claim <task-id> # Claim a task wg log <task-id> "msg" # Log progress wg done <task-id> # Mark complete
Quick reference
| Command | Purpose |
|---|---|
wg quickstart | Orient yourself at session start |
wg service start | Start the coordinator service |
wg service status | Check if service is running |
wg add "Title" -d "Desc" | Create a task |
wg add "X" --blocked-by Y | Create task with dependency |
wg list | All tasks with status |
wg ready | Tasks available to work on |
wg show <id> | Full task details |
wg context <id> | Inputs from dependencies |
wg agents | Running agents |
wg done <id> | Complete a task |
wg submit <id> | Submit verified task for review |
wg fail <id> --reason "why" | Mark task failed |
wg log <id> "msg" | Log progress |
wg artifact <id> <path> | Record output file |
wg impact <id> | What depends on this? |
wg analyze | Full health report |
All commands support --json for structured output. Run wg --help for the full command list.