Skip to content
BMO

Staged Workflow

Staged workflow separates planning from execution. The agent first produces a plan artifact and waits. No file writes are permitted until the plan is approved.

Staged workflow plan review in BMO
A BMO TUI capture-style artifact showing the staged plan gate before execution is unlocked.

The agent stores the plan with set_plan_artifact. There is no submit_plan tool. In the TUI, the plan appears in the dedicated Plan panel and the status bar.

[options.staged_workflow]
enabled = true
require_approval_before_execute = true

How it works

Section titled “How it works”
  1. Agent receives your request and enters plan phase
  2. Agent produces a structured plan (no writes allowed)
  3. You review the plan in the TUI
  4. Type /approve to advance to execute phase — agent carries out the plan
  5. Or type /plan to send the agent back to revise

With require_approval_before_execute: false, the agent moves to execute automatically once a plan exists (useful for headless/autopilot runs).

Maintained review prompts

Section titled “Maintained review prompts”

Maintained review prompts are workspace-meta artifacts at the agent-cli workspace root (docs/prompts/), not inside the BMO git tree. From a checkout of agent-cli/bmo/, browse them at ../docs/prompts/. Load bodies with view (pointing at the workspace path) or open the file in the TUI, then submit the text under Execute intent. Use /intent discuss … to explore a prompt without running mutating tools. Commands and recipes still use /run, run_recipe, or run_custom_command. Note: the prompt corpus is not part of the published BMO repository — it lives in the maintainer’s local agent-cli/ workspace.

Options

Section titled “Options”
OptionDefaultDescription
enabledfalseEnable staged workflow
require_approval_before_executetrueBlock execution until /approve is typed

Slash commands

Section titled “Slash commands”
CommandDescription
/approveApprove the current plan and advance to execute
/planReturn to plan phase to revise

When to use

Section titled “When to use”

Use staged workflow when you want to review what the agent intends to do before it touches any files. Good for:

  • Large refactors
  • Multi-file feature additions
  • Any task where reviewing intent upfront saves debugging later

Architect / editor mode

Section titled “Architect / editor mode”

For different models for plan vs execute (similar to other tools’ “architect mode”), enable Phase model routing alongside staged workflow: [options.phase_model] sets planning_model and execution_model. Copy-paste preset: Architect / editor preset.

Section titled “Related”