SDD Observatory
Stars
50.7K
Contributors
154
Open issues
99
Age
1.3 years
Last push
7 hours ago
Latest release
v6.10.0 Jul 3, 2026
Week of Jul 19, 2025: 6 commits Week of Jul 26, 2025: 9 commits Week of Aug 2, 2025: 17 commits Week of Aug 9, 2025: 43 commits Week of Aug 16, 2025: 12 commits Week of Aug 23, 2025: 3 commits Week of Aug 30, 2025: 18 commits Week of Sep 6, 2025: 6 commits Week of Sep 13, 2025: 1 commit Week of Sep 20, 2025: 0 commits Week of Sep 27, 2025: 52 commits Week of Oct 4, 2025: 12 commits Week of Oct 11, 2025: 41 commits Week of Oct 18, 2025: 29 commits Week of Oct 25, 2025: 49 commits Week of Nov 1, 2025: 32 commits Week of Nov 8, 2025: 31 commits Week of Nov 15, 2025: 30 commits Week of Nov 22, 2025: 27 commits Week of Nov 29, 2025: 46 commits Week of Dec 6, 2025: 58 commits Week of Dec 13, 2025: 38 commits Week of Dec 20, 2025: 32 commits Week of Dec 27, 2025: 25 commits Week of Jan 3, 2026: 16 commits Week of Jan 10, 2026: 39 commits Week of Jan 17, 2026: 43 commits Week of Jan 24, 2026: 47 commits Week of Jan 31, 2026: 51 commits Week of Feb 7, 2026: 22 commits Week of Feb 14, 2026: 30 commits Week of Feb 21, 2026: 25 commits Week of Feb 28, 2026: 18 commits Week of Mar 7, 2026: 117 commits Week of Mar 14, 2026: 78 commits Week of Mar 21, 2026: 38 commits Week of Mar 28, 2026: 23 commits Week of Apr 4, 2026: 27 commits Week of Apr 11, 2026: 9 commits Week of Apr 18, 2026: 20 commits Week of Apr 25, 2026: 23 commits Week of May 2, 2026: 5 commits Week of May 9, 2026: 4 commits Week of May 16, 2026: 7 commits Week of May 23, 2026: 21 commits Week of May 30, 2026: 2 commits Week of Jun 6, 2026: 9 commits Week of Jun 13, 2026: 10 commits Week of Jun 20, 2026: 24 commits Week of Jun 27, 2026: 11 commits Week of Jul 4, 2026: 12 commits Week of Jul 11, 2026: 5 commits
Commits per week, last 12 months · 1.4K total

Metrics updated Jul 18, 2026 · collected automatically from the GitHub API

Core approach

Simulates an agile team of 12+ specialized AI agent personas (PM, Architect, Developer, UX, and more) that facilitate structured workflows across a four-phase lifecycle. Planning depth auto-adjusts to project complexity, from bug fixes to enterprise systems. Installed via npx as a module ecosystem with 34+ workflows in the core module.

Workflow

  1. Phase 1 · Analysis (optional) — brainstorming, idea forging, domain/market/technical research, product brief
  2. Phase 2 · Planning — PRD and UX workflows
  3. Phase 3 · Solutioning — architecture, epic and story creation, implementation-readiness gate
  4. Phase 4 · Implementation — sprint planning, story development, code review, course correction, retrospective
  5. Parallel Quick Flow track — quick-dev and an unattended dev loop for small changes

Supported tools

Strengths

  • Broadest lifecycle coverage of the frameworks tracked here — ideation through retrospective, not just spec-to-code
  • Modular ecosystem (builder, test architect, game dev, creative suite) plus web bundles that offload planning to flat-rate LLM subscriptions
  • Scale-adaptive planning depth, with a help workflow that guides users to the next step
  • Free and open source with a large, active community and frequent releases

Limitations

  • Heavyweight and ceremony-rich compared to lighter SDD kits — many agents, phases, and workflows to learn
  • V6 was a breaking restructure; upgrading from v4/v5 requires a dedicated migration guide
  • Prerequisites span Node.js 20.12+, Python 3.10+, and uv
  • The repo does not visibly dogfood its own story/PRD artifacts in-tree, and BMAD/BMad are trademarks with usage restrictions

Notes

BMAD reads less like a spec format and more like an operating model: it recreates the roles of an agile team as agent personas and routes work through their ceremonies. That makes it the most complete lifecycle of the tracked frameworks and also the heaviest. Its repository moved from bmadcode/BMAD-METHOD to the bmad-code-org organization, and GitHub reports its license as “Other” because the MIT text carries a custom preamble. Well-starred active adopters are hard to observe in public — code search finds ~400 v4-era .bmad-core configs, but the most visible ones have gone idle — so evidence of long-term outcomes is currently thin.

Added Jul 18, 2026 · Assessment last reviewed Jul 18, 2026 · How we track