blueclover22/mw_agent_kit
Spec-driven development workflow for Claude Code — built around an approved design doc, where each stage waits for your approval.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code $ npx degit blueclover22/mw_agent_kit/agents ./rig-mw_agent_kit/agents $ npx degit blueclover22/mw_agent_kit/skills ./rig-mw_agent_kit/skills
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit blueclover22/mw_agent_kit/skills/brainstorming .claude/skills/brainstorming $ npx degit blueclover22/mw_agent_kit/skills/commit .claude/skills/commit $ npx degit blueclover22/mw_agent_kit/skills/design-doc-template .claude/skills/design-doc-template $ npx degit blueclover22/mw_agent_kit/skills/dev-kickoff .claude/skills/dev-kickoff $ npx degit blueclover22/mw_agent_kit/skills/dev-resume .claude/skills/dev-resume $ npx degit blueclover22/mw_agent_kit/skills/doc-audit .claude/skills/doc-audit $ npx degit blueclover22/mw_agent_kit/skills/reverse-engineering .claude/skills/reverse-engineering $ npx degit blueclover22/mw_agent_kit/skills/review-report .claude/skills/review-report $ npx degit blueclover22/mw_agent_kit/skills/roadmap-planning .claude/skills/roadmap-planning $ npx degit blueclover22/mw_agent_kit/skills/setup .claude/skills/setup $ npx degit blueclover22/mw_agent_kit/skills/teardown .claude/skills/teardown $ npx degit blueclover22/mw_agent_kit/skills/verify-checklist .claude/skills/verify-checklist
$ curl -fsSL --create-dirs -o .claude/agents/analyzer.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/analyzer.md $ curl -fsSL --create-dirs -o .claude/agents/auditor.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/auditor.md $ curl -fsSL --create-dirs -o .claude/agents/coder.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/coder.md $ curl -fsSL --create-dirs -o .claude/agents/doc-editor.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/doc-editor.md $ curl -fsSL --create-dirs -o .claude/agents/planner.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/planner.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/researcher.md $ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/blueclover22/mw_agent_kit/main/agents/reviewer.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 7,204 rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(rm -rf *)",
"Read(**/*.pem)",
"Bash(rm -rf /)",
"Bash(git push --force:*)",
"Bash(sudo *)",
"Read(.env)",
"Bash(rm -rf /*)",
"Read(./.env.*)",
"Read(~/.aws/**)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(sudo:*)",
"Bash(git reset --hard*)",
"Bash(git reset --hard:*)",
"Read(.env.*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(rm:*)",
"Bash(git rebase *)",
"Bash(wget *)",
"Bash(npm publish:*)",
"Bash(git commit *)",
"Bash(gh pr merge *)"
]
}
} Skills (12)
brainstormingcommitdesign-doc-templatedev-kickoffdev-resumedoc-auditreverse-engineeringreview-reportroadmap-planningsetupteardownverify-checklist
Subagents (7)
| analyzer model: opus | Use during mak:reverse-engineering to read source and fill the assigned copied docs/ set with observed facts, or on explicit request for a standalone codebase-analysis or code-vs-doc report. Non-inter |
| auditor model: opus | Use for mak:doc-audit cross-document consistency checks across design docs, the roadmap, and permanent docs, including cited path/symbol existence. Report-only; never edits audited docs. Not for imple |
| coder model: inherit | Use to implement code changes the user asked for. Trivial/Small fixes (simple bugs, typos, independent single-file changes) need no approved plan; Standard/Risky work requires an approved design doc. |
| doc-editor model: sonnet | Use to edit, organize, or sync Markdown (docs/, design docs, README, CHANGELOG) on explicit request or confirmed post-feature doc sync, or to create one whose path and content the main thread specifie |
| planner model: opus | Use only when explicitly delegated architecture analysis for Standard/Risky work. Read-only — investigates assigned code/docs and returns an Architecture Brief with options, recommendation, risks, and |
| researcher model: opus | Use to research a topic from external sources and write the findings as a document at a path the main thread specifies. Non-interactive; writes documents, never source code. Not for architecture consu |
| reviewer model: opus | Use after implementation or mak:coder when code needs review against its design doc/conventions, or on explicit requests ("code review", "PR review", "review the implementation"). Report-only; never m |
Similar rigs
juzser/blacksmith
Autonomous agent factory — loop runner, worktree engine, and a local dashboard over the event log.
Orchestrator 2.6k tok ·
OMOCHInoHOSHI/dotfiles-claude
.claude
Pragmatist 293 tok ·
BillDX/spicy-to-dark-factory-starter
Workshop starter for 'From Spicy Autocomplete to Dark Factory: Agentic Coding in Practice' — a FastAPI mini-app with a full Claude Code harness (agents, slash commands, hooks, sandbox settings) and self-paced exercises.
Orchestrator 706 tok ·
homeofe/improvements
Portable, LLM-agnostic framework for multi-model AI workflows: model routing, structured agent handoff (AAHP), and a phased agent pipeline. MIT.
Orchestrator 1.4k tok ·