Synopsis

aigon research-autopilot <research-id> [agents...] [--eval-agent=<agent>]
aigon research-autopilot status <research-id>
aigon research-autopilot stop <research-id>

Shortcuts

• /arap (slash command) · aigon arap (CLI)

Description

Runs a full research cycle hands-off: starts all agents, waits for findings, then spawns an evaluator to synthesise results and recommend features. The equivalent of running research-start + research-eval automatically.

Usage

# Fleet research — two agents, auto-eval when done
aigon research-autopilot 23 cc ag --eval-agent=cc
 
# Check AutoConductor status
aigon research-autopilot status 23
 
# Stop the AutoConductor (agents continue independently)
aigon research-autopilot stop 23

From the dashboard

Click Research Autopilot on any backlog research card.

What it does

1. Calls research-start to set up agent worktrees and launch all agents
2. Polls workflow state every 30 seconds
3. When all agents signal research-complete, spawns the evaluator
4. Evaluator synthesises findings and recommends features
5. AutoConductor exits with a summary

Related

• research-start — manual research setup
• research-eval — manual evaluation
• Research Workflow guide