BMAD-METHOD/.patch/821/integration-options.md

1.7 KiB
Raw Blame History

Integration Options for PR #821

Date: 2025-10-28 Branch: pr-821-review

Option 1: Accept as Static "Snapshot" Module

  • Location: extras/subagentic/ or src/modules/subagentic/ (doc-only)
  • Scope: Keep Claude/OpenCode content intact; mark as snapshot
  • Work: Clean whitespace, add README, wire docs
  • Effort: 46 hours
  • Risks: Drift from BMAD, dual system confusion
  • Scope: Extract patterns from subagent definitions (principles, commands)
  • Work:
    1. Spike adapter for 12 agents → *.agent.yaml
    2. Design templates to generate MD (if needed) for editor consumption
    3. Integrate via bmad CLI workflows and sidecars
  • Effort: 1016 hours (spike + template wiring)
  • Benefits: Single source of truth; aligned with v6 vision

Option 3: Reference as External Alternative

  • Scope: Do not merge code; link to agentic-toolkit in docs
  • Work: Update README.md + docs under docs/installers-bundlers/
  • Effort: 12 hours
  • Benefits: Zero maintenance; users can opt-in

Option 4: Hybrid

  • Scope: Accept 3 simple workflow agents (Create PRD, Generate Tasks, Process Task List)
  • Work: Convert 3 agents to BMAD YAML; ship as src/modules/simple/
  • Effort: 610 hours
  • Benefits: Adds unique value with minimal footprint

Decision Criteria

  • Architectural alignment with v6 (dynamic, template-driven)
  • Maintenance burden and ownership
  • User clarity and documentation needs
  • Ability to validate/test within existing tooling

Next Steps

  • If Option 2 or 4: Build a small adapter prototype for one agent (e.g., master)
  • If Option 3: Add docs and close PR with appreciation and cross-link
  • Prepare PR comment summarizing analysis and recommendation