/config

Maestro Config

View and edit Maestro configuration for this project. All settings are stored in .maestro/config.yaml.

No Arguments or show — Display Current Config

Read .maestro/config.yaml and display it with inline annotations:

+---------------------------------------------+
| Maestro Configuration                       |
+---------------------------------------------+

  Execution:
    default_mode         checkpoint
    default_model        sonnet

  Quality Gates:
    max_qa_iterations    5          (QA review cycles per story)
    max_self_heal        3          (auto-fix attempts per story)
    run_tsc              true       (TypeScript type checking)
    run_lint             true       (linter)
    run_tests            true       (test suite)

  Cost Tracking:
    enabled              true
    forecast             true       (estimate before starting)
    ledger               true       (log per-story costs)
    budget_enforcement   true       (pause at budget limit)

  Integrations:
    Kanban:
      provider           (not set)
      sync_enabled       false
    Knowledge Base:
      provider           (not set)
      vault_path         (not set)
      sync_enabled       false

  (i) Use /maestro config set KEY VALUE to change settings.
  (i) Use /maestro config reset to restore defaults.

If .maestro/config.yaml does not exist:

[maestro] Config not found. Run /maestro init first.

set KEY VALUE — Update a Setting

Support dot-notation for nested keys. Validate both key and value.

Valid Keys and Values

Key	Valid Values	Description
default_mode	yolo, checkpoint, careful	Execution mode
default_model	haiku, sonnet, opus	Default implementation model
quality.max_qa_iterations	1-10 (integer)	Max QA review cycles
quality.max_self_heal	1-5 (integer)	Max auto-fix attempts
quality.run_tsc	true, false	Run TypeScript checks
quality.run_lint	true, false	Run linter
quality.run_tests	true, false	Run tests
cost_tracking.enabled	true, false	Enable cost tracking
cost_tracking.forecast	true, false	Show forecast before start
cost_tracking.ledger	true, false	Log per-story costs
cost_tracking.budget_enforcement	true, false	Pause at budget limit
integrations.kanban.provider	asana, jira, linear, github, null	Kanban provider
integrations.kanban.sync_enabled	true, false	Auto-sync stories
integrations.kanban.project_id	(string)	Project/board ID
integrations.knowledge_base.provider	obsidian, notion, null	Knowledge base provider
integrations.knowledge_base.vault_path	(path string)	Path to vault/workspace
integrations.knowledge_base.sync_enabled	true, false	Auto-sync enabled
models.planning	haiku, sonnet, opus	Model for decomposition/architecture
models.execution	haiku, sonnet, opus	Model for story implementation
models.review	haiku, sonnet, opus	Model for QA review
models.simple	haiku, sonnet, opus	Model for fix agents
models.research	haiku, sonnet, opus	Model for web research
scheduler.enabled	true, false	Enable cron-based scheduling

Validation Rules

1. Reject unknown keys with a helpful message listing valid keys.
2. Reject invalid values with the list of allowed values for that key.
3. For integration providers, check that the corresponding MCP server or CLI tool is available. Warn if not detected but still allow setting (user may install it later).
4. For vault_path, verify the path exists if it's set.

Output on Success

[maestro] Updated: default_mode = yolo

  (i) This affects all future Maestro sessions on this project.

Output on Integration Provider Change

When setting a kanban or knowledge base provider, run a quick connectivity check:

[maestro] Updated: integrations.kanban.provider = asana

  Connectivity check:
    (ok) Asana MCP server detected
    (!)  No project_id configured yet.
         Set it with: /maestro config set integrations.kanban.project_id YOUR_PROJECT_ID

  (i) Enable auto-sync: /maestro config set integrations.kanban.sync_enabled true

reset — Restore Defaults

Ask for confirmation before resetting:

Use AskUserQuestion:

• Question: "Reset all configuration to defaults? Integration settings will be cleared."
• Header: "Reset"
• Options:
  1. label: "Yes, reset everything", description: "Overwrite config.yaml with default values"
  2. label: "Cancel", description: "Keep current configuration"

On confirmation, regenerate config.yaml with the default template (same as maestro init generates).

Interactive Mode (no arguments)

When $ARGUMENTS is empty, after showing the current config, use AskUserQuestion to offer an interactive menu:

Question: "What would you like to configure?"

Options:

1. Execution mode — "Change default mode (currently: [mode])"
2. Quality gates — "Toggle tsc, lint, tests, adjust QA/self-heal limits"
3. Integrations — "Configure kanban or knowledge base providers"
4. Cost tracking — "Toggle forecast, ledger, budget enforcement"

If "Execution mode":

Use AskUserQuestion:

Question: "Default execution mode?"

Options:

1. yolo — "Auto-approve everything. Maximum speed, minimum oversight"
2. checkpoint — "Pause after each story for review (recommended)"
3. careful — "Pause after each phase for granular control"

Update default_mode in config.

If "Quality gates":

Use AskUserQuestion with multiSelect:

Question: "Which quality checks should run during self-heal? (select all that apply)"

Options (multiSelect: true):

1. TypeScript (tsc) — "Run tsc --noEmit to catch type errors"
2. Linter — "Run the project linter (eslint, etc.)"
3. Tests — "Run the test suite"

Update quality.run_tsc, quality.run_lint, quality.run_tests accordingly.

Then use AskUserQuestion:

Question: "Adjust QA and self-heal limits?"

Options:

1. Keep current — "QA: [max_qa] iterations, Self-heal: [max_self_heal] attempts"
2. Strict — "QA: 3 iterations, Self-heal: 2 attempts (fail faster)"
3. Patient — "QA: 8 iterations, Self-heal: 5 attempts (try harder)"

If "Integrations":

Use AskUserQuestion:

Question: "Which integration to configure?"

Options:

1. Kanban provider — "Sync stories with project management (currently: [provider or 'none'])"
2. Knowledge base — "Connect second brain (currently: [provider or 'none'])"
3. Back — "Return to main menu"

If Kanban selected, use AskUserQuestion:

Question: "Kanban provider?"

Options:

1. GitHub Issues — "Uses gh CLI. No extra setup needed"
2. Asana — "Requires Asana MCP Server"
3. Jira — "Requires Atlassian MCP Server"
4. Linear — "Requires Linear MCP Server"

Update config and run connectivity check.

If Knowledge base selected, suggest running /maestro brain connect for the full guided setup.

If "Cost tracking":

Use AskUserQuestion with multiSelect:

Question: "Cost tracking settings? (select all to enable)"

Options (multiSelect: true):

1. Show forecast before starting — "Estimate cost before execution begins"
2. Log per-story costs — "Track token usage in token-ledger.md"
3. Enforce budget limits — "Pause execution when budget is reached"

Update cost_tracking settings accordingly.

After any change, show confirmation and ask if they want to configure something else.