A community-driven registry for Claude, Cursor, Windsurf, Cline & more. Not affiliated with Anthropic.
Are you the author? Sign in to claim
Claude Code plugins for spec-driven development workflows: SPECLAN integration, BMAD-METHOD interop, and SDD automation
Claude Code plugins for specification-driven development, project governance, and task management.
This repo is intended to be cloned and customized to fit your software development process. Fork it, adjust the plugins to match your workflows, and use it as a starting point rather than a fixed dependency.
Claude Code helper for the SPECLAN VSCode extension - manage project specifications as interlinked markdown files.
Using BMAD? See the BMAD Integration Guide for a full walkthrough of the round-trip workflow between BMAD and SPECLAN.
| Flow | Description |
|---|---|
| Ask | Query specs with full context — hierarchy, relationships, status |
| Grill | Interview the user question-by-question to harden specs before approval |
| Impl → Spec | Sync session work to SPECLAN specifications |
| Spec → Impl | Plan, implement, and review approved specs (manual three-phase workflow) |
| BMAD → Speclan | Convert BMAD planning artifacts (PRD, Epics) to SPECLAN format |
| Speclan → BMAD | Generate BMAD-standard PRD from approved specs (two approaches: BMAD-native or standalone converter) |
| Speckit → Speclan | Convert speckit specifications to SPECLAN format |
| Speclan → Speckit | Inject approved SPECLAN specs into speckit commands |
| Command | Description |
|---|---|
/speclan:ask | Ask about any SPECLAN entity — get holistic context with hierarchy, relationships, and status |
/speclan:grill | Grill the user on a spec — one-question-at-a-time interview that reduces ambiguities, clarifies contradictions, and closes definition gaps before approval (--language for non-English sessions) |
/speclan:status | Check project setup, installed plugins, and spec statistics |
/speclan:sync | Sync session work to SPECLAN specs - identifies implemented features |
/speclan:plan-manual | Create a manual implementation plan from approved specs |
/speclan:implement-manual | Implement the next pending item from a manual plan |
/speclan:review-manual | Review all in-review items, verify spec compliance and code quality |
/speclan:from-bmad | Convert BMAD planning artifacts (PRD, Epics, Architecture) to SPECLAN format |
/speclan:to-bmad | Generate BMAD-standard PRD using BMAD's native PRD creation workflow |
/speclan:to-bmad-prd | Generate BMAD-standard PRD using standalone converter (no BMAD skills required) |
/speclan:from-speckit | Convert speckit specs to SPECLAN format |
/speclan:to-speckit | Inject approved SPECLAN Features into speckit |
| Agent | Description |
|---|---|
feature-verifier | Deep 5-layer verification of feature implementation against specs |
requirement-verifier | Verify individual requirements are satisfied by code |
implementation-order | Analyze feature dependencies and determine optimal build order |
bmad-converter | Bidirectional conversion between BMAD-METHOD and SPECLAN formats |
speckit-converter | Bidirectional conversion between speckit and SPECLAN formats |
code-hygiene-reviewer | Review code structure, naming, duplication, dead code, type safety |
architecture-reviewer | Review design patterns, SOLID principles, coupling, module boundaries |
| Skill | Description |
|---|---|
speclan-format | Foundational knowledge for all speclan/ file operations |
speclan-query | Query specs by type, status, or parent relationship (JSON output) |
speclan-id-generator | Collision-free random ID generation for SPECLAN entities |
ask | Answer spec questions with full context — hierarchy, relationships, status |
grill | Interview workflow that questions a spec's ambiguities, contradictions, and gaps, applying agreed fixes to the spec |
sync-from-session | Workflow for capturing session work as SPECLAN specs |
plan-manual | Create manual implementation plans from approved specs |
implement-manual | Implement next item from a manual plan (plan → implement → review) |
review-manual | Review all in-review items, verify spec compliance and code quality |
| Event | Behavior |
|---|---|
SessionStart | Detects speclan/ directory, counts specs, checks plugin dependencies |
PreToolUse (Write/Edit) | Guards locked specs - blocks direct edits to in-development/released/deprecated entities, rejects invalid status values |
Three-phase workflow for implementing approved specs with human oversight at each stage:
/speclan:plan-manual → /speclan:implement-manual → /speclan:review-manual
1. Plan — /speclan:plan-manual queries all approved specs and generates a plan file at speclan/.local/plans/. The plan is a pure-data checklist with checkbox states tracking progress. Only approved items and their ancestor chain are included.
2. Implement — /speclan:implement-manual picks up the plan file and implements the next [ ] item. Run repeatedly — each invocation implements one feature (with all its requirements) or one individual item. Checkboxes transition: [ ] → [~] → [?].
3. Review — /speclan:review-manual reviews ALL [?] items at once, verifying spec compliance (acceptance criteria) and code quality (hygiene + architecture subagents). On acceptance, checkboxes transition [?] → [x].
Plan file checkboxes: [ ] pending → [~] in-dev → [?] in-review → [x] done
Spec statuses: approved → in-development → under-test (unchanged by review)
Spec files are read-only during review — status transitions to released happen after human testing.
Goal (G-###)
└── Feature (F-####) [tree structure, directory-based]
└── Requirement (R-####) [directory-based, acceptance criteria inline]
Plus: ChangeRequest (CR-####) for modifying locked entities, Template (UUID) for spec scaffolding.
speclan/
├── goals/ # G-### Strategic goals (flat files)
├── features/ # F-#### Feature tree (hierarchical directories)
│ └── F-1234-my-feature/
│ ├── F-1234-my-feature.md
│ ├── requirements/
│ │ └── R-5678-some-req/
│ └── change-requests/
├── requirements/ # R-#### Top-level requirements
└── templates/ # Specification templates
draft → review → approved → in-development → under-test → released → deprecated
Specs in draft, review, approved are directly editable. Locked statuses (in-development, under-test, released) require a Change Request. deprecated is permanently frozen.
Manage project constitution files - rules, best practices, patterns, and guardrails for LLM/coding agents.
| Command | Description |
|---|---|
/constitution:check | Check code or session work for constitution violations |
/constitution:add | Add rules, best practices, or guidelines to the constitution |
/constitution:clarify | Clarify or refine existing constitution rules |
| Skill | Description |
|---|---|
constitution-format | Foundational knowledge for constitution file structure and conventions |
Filesystem-based task management with git integration - track tasks through open/wip/done lifecycle.
| Command | Description |
|---|---|
/todo:add | Create a new task in the todo system |
/todo:start | Start working on a task - move to wip, create branch, assign |
/todo:update | Log implementation progress - auto-detect work from git diff |
/todo:done | Mark a task as complete - move to done, finalize log |
| Skill | Description |
|---|---|
todo-format | Foundational knowledge for todo file structure and lifecycle |
| Event | Behavior |
|---|---|
SessionStart | Detects todo/ directory and provides task context |
PreToolUse (Read/Write/Edit) | Auto-injects todo format knowledge when accessing task files |
todo/
├── open/ # Pending tasks
├── wip/ # Work in progress
└── done/ # Completed tasks
| Plugin | Requires |
|---|---|
| speclan | SPECLAN directory (speclan/) |
| constitution | Constitution file (constitution.md) - created automatically |
| todo | Todo directory (todo/) - created automatically |
| Dependency | Required By |
|---|---|
| BMAD-METHOD | /speclan:from-bmad, /speclan:to-bmad, /speclan:to-bmad-prd |
| speckit plugin | /speclan:from-speckit, /speclan:to-speckit |
MIT
Native macOS app to monitor Claude AI usage limits and watch your coding sessions live
npx CLI installing 100+ agents, commands, hooks, and integrations in one command
干净、强大、属于你的 AI Agent 平台 --AI agents, without the clutter.
Pocket Flow: Codebase to Tutorial