Guides

Policy & governance

Use ReactorAI project policy, quality gates, and role-based approvals to ensure refactor output meets release controls before publication.

Time: 11 minAudience: Platform admins and engineering managers

Governance model

Publish mode: choose pr_only or controlled_push per project.
Required approver roles: gate publishing on explicit role approvals.
Protected branches: prevent direct publish to sensitive branches.
Quality gates: block publish when gate status is not passed.

Set project policy in ReactorAI

Set policy defaults in the project settings page. Keep policy changes versioned and communicate rollout windows to affected teams.

Recommended governance baseline

# Recommended policy baseline in ReactorAI
publish_mode: pr_only
required_approver_roles:
  - manager
protected_branches:
  - main
  - release
audit_export_enabled: true

Approve and publish safely

Keep PR publish mode as the default. Use direct publish only where explicit policy allows it and protected branch checks are satisfied.

Run approval flow in dashboard

# In ReactorAI run details:
# 1) Review findings and patch groups
# 2) Confirm gate status is passed
# 3) Capture approver comments
# 4) Approve or reject according to release criteria

# Keep direct publish disabled for protected branches by default

Audit and compliance evidence

Maintain recurring governance reviews so your controls stay effective as projects and teams scale.

Governance reporting cadence

# Governance evidence checklist
- Export approval history weekly
- Track blocked publish reasons monthly
- Review policy drift across projects each sprint