jbeat30/claude-subagent-workflow
A Claude code plugin for sub-agent workflows led by the project manager, from planning and implementation to independent QA. Become a small one-person business
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 jbeat30/claude-subagent-workflow/agents ./rig-claude-subagent-workflow/agents $ npx degit jbeat30/claude-subagent-workflow/skills ./rig-claude-subagent-workflow/skills $ npx degit jbeat30/claude-subagent-workflow/hooks ./rig-claude-subagent-workflow/hooks
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 jbeat30/claude-subagent-workflow/skills/add-role .claude/skills/add-role $ npx degit jbeat30/claude-subagent-workflow/skills/check .claude/skills/check $ npx degit jbeat30/claude-subagent-workflow/skills/run .claude/skills/run $ npx degit jbeat30/claude-subagent-workflow/skills/setting .claude/skills/setting $ npx degit jbeat30/claude-subagent-workflow/skills/setup .claude/skills/setup
$ curl -fsSL --create-dirs -o .claude/agents/api-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/api-agent.md $ curl -fsSL --create-dirs -o .claude/agents/architecture-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/architecture-agent.md $ curl -fsSL --create-dirs -o .claude/agents/backend-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/backend-agent.md $ curl -fsSL --create-dirs -o .claude/agents/dev-lead.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/dev-lead.md $ curl -fsSL --create-dirs -o .claude/agents/frontend-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/frontend-agent.md $ curl -fsSL --create-dirs -o .claude/agents/planner-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/planner-agent.md $ curl -fsSL --create-dirs -o .claude/agents/planning-lead.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/planning-lead.md $ curl -fsSL --create-dirs -o .claude/agents/qa-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/qa-agent.md $ curl -fsSL --create-dirs -o .claude/agents/test-agent.md https://raw.githubusercontent.com/jbeat30/claude-subagent-workflow/main/agents/test-agent.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Skills (5)
Subagents (9)
| api-agent model: sonnet | Use when a shared API contract, shared types, or client integration that both server and screens depend on must change. Implement within the assigned contract and client files only. |
| architecture-agent model: sonnet | Use when a change may affect callers, shared modules, API contracts, data, security, or regressions, and the impact must be analyzed before planning. Read-only analysis that returns a report. |
| backend-agent model: sonnet | Use when server routes, business logic, existing authentication handling, or data processing must change. Implement within the assigned backend files only and report contract impacts. |
| dev-lead model: sonnet | Reserved for the leader organization planned for v0.8.0; the PM does not call it in v0.7.0. Implements a confirmed plan: splits the plan into sequential work, calls the assigned implementers one at a |
| frontend-agent model: sonnet | Use when screens, markup, accessibility, or UI behavior must change. Implement within the assigned frontend files only and report contract or scope impacts. |
| planner-agent model: sonnet | Use when a request needs requirement breakdown, discovery of items the user did not state, assumptions, and verifiable acceptance criteria before planning. Read-only analysis that returns a report. |
| planning-lead model: sonnet | Reserved for the leader organization planned for v0.8.0; the PM does not call it in v0.7.0. Coordinates planning for an L2 or L3 request: runs the planner, architecture, and plan QA analysts, integrat |
| qa-agent model: sonnet | Use for independent verification in two modes. Plan mode reviews a plan for completeness against the original request. Final mode verifies changes, check records, and the completion checklist against |
| test-agent model: sonnet | Use when unit, UI, API, or end-to-end tests must be written for acceptance criteria or regressions. Write tests within the assigned test files only and request check runs from the PM. |
Similar rigs
quboqin/template-agenticide-vibecoding
🤖 Claude Code定制化研发流程系统 - 基于Spec-Driven和Test-Driven的AI智能体团队协作框架。构建AI优先的软件开发新范式,让AI智能体团队替代传统人类开发团队。
Orchestrator 1.6k tok · · Q0
xahdiyar37-hub/template-agenticide-vibecoding
Claude Code agentic AI development workflow template - 7 AI Agents, spec-driven + test-driven development
Orchestrator 1.6k tok · · Q0
gorock007/jarvis
Jarvis: one lead agent you talk to, a fleet of workers that builds. Runs in Superset or tmux, with Claude Code, Codex, OpenCode or any agent CLI as workers.
Minimalist 167 tok · · Q15
OlegTestov/prove-it-mini
A pytest Stop hook for Claude Code: two repair attempts, then an explicit unverified result. No model calls, MIT.
Minimalist 332 tok · · Q3