Dev-Toolbelt/dev-team-agents
Multi-agent development harness for Claude Code, Codex CLI, and opencode — stack-agnostic, project-aware agents, commands, skills, and hooks for planning, building, reviewing, and documenting software.
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 Dev-Toolbelt/dev-team-agents/.claude ./rig-dev-team-agents # inspect, then merge into .claude/ 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 Dev-Toolbelt/dev-team-agents/.claude/skills/agent-creator .claude/skills/agent-creator $ npx degit Dev-Toolbelt/dev-team-agents/.claude/skills/repo-release-prep .claude/skills/repo-release-prep $ npx degit Dev-Toolbelt/dev-team-agents/skills/skill-creator .claude/skills/skill-creator
$ curl -fsSL --create-dirs -o .claude/agents/backend-developer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/backend-developer.md $ curl -fsSL --create-dirs -o .claude/agents/backend-reviewer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/backend-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/backend-test-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/backend-test-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/database-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/database-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/devops-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/devops-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/frontend-developer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/frontend-developer.md $ curl -fsSL --create-dirs -o .claude/agents/frontend-reviewer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/frontend-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/frontend-test-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/frontend-test-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/mobile-developer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/mobile-developer.md $ curl -fsSL --create-dirs -o .claude/agents/product-analyst.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/product-analyst.md $ curl -fsSL --create-dirs -o .claude/agents/qa-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/qa-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/security-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/security-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/seo-specialist.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/seo-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/setup-assistant.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/setup-assistant.md $ curl -fsSL --create-dirs -o .claude/agents/software-architect.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/software-architect.md $ curl -fsSL --create-dirs -o .claude/agents/technical-writer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/technical-writer.md $ curl -fsSL --create-dirs -o .claude/agents/ui-ux-designer.md https://raw.githubusercontent.com/Dev-Toolbelt/dev-team-agents/main/agents/ui-ux-designer.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": ".*",
"hooks": [
{
"type": "command",
"command": "scripts/hooks/pre-tool-use.sh"
}
]
}
]
}
} Skills (3)
Subagents (18)
| backend-developer model: sonnet | Implements backend features following the project's architecture and code standards. Adapts to whatever server-side style the project uses — API-first or server-rendered, layered or flat. Writes natur |
| backend-reviewer model: sonnet | Specialized code reviewer for backend changes. Covers API contracts, database transactions, N+1 queries, auth/authz, background jobs, race conditions, SOLID, DI, and security. Invoked by the review-ro |
| backend-test-specialist model: sonnet | Creates backend tests (unit, integration, E2E) for code that was written or modified. Only activates when the project has a test culture (CLAUDE.md indicates tests are required, or user explicitly req |
| code-reviewer model: sonnet | Reviews code for quality, correctness, security, and standards compliance. Covers: design patterns, SOLID, Object Calisthenics, DRY, code repetition, race conditions, silent bugs, linting, readability |
| database-specialist model: sonnet | Expert in database design, query optimization, indexing strategy, and schema decisions across relational, document, key-value, and column-family engines, whether self-hosted or managed in the cloud. U |
| devops-specialist model: sonnet | Infrastructure specialist. Sets up dev and production environments, provisions servers, configures CI/CD pipelines, deploys to cloud or self-hosted infra in a cost-optimized way, and manages monitorin |
| frontend-developer model: sonnet | Implements frontend features following the project's design system and architecture. Works in both decoupled SPAs (React, Vue, Svelte, Angular) and server-rendered templates (Blade, Twig, ERB, Jinja). |
| frontend-reviewer model: sonnet | Specialized code reviewer for frontend changes. Covers component design, re-renders, accessibility, bundle size, state management, XSS, loading states, error boundaries, CSS quality, and type safety. |
| frontend-test-specialist model: sonnet | Creates frontend tests (component, integration, E2E) for UI code written or modified. Only activates when the project has a test culture or user explicitly requests tests. Covers component testing, us |
| mobile-developer model: sonnet | Implements mobile features for iOS and Android, whether the project is native or cross-platform. Detects the project's mobile stack and follows its platform conventions. Use for any mobile implementat |
| product-analyst model: opus | The lead agent for planning. Reads a request or requirements text, interrogates it against a fixed set of lenses, runs a short focused conversation to close the open decisions, and produces a purely b |
| qa-specialist model: sonnet | Validates product behavior, user flows, and regression risk from a quality assurance perspective. Focuses on what the product does (not how the code does it). Tests critical paths, edge cases, and int |
| security-specialist model: opus | Performs security reviews covering OWASP Top 10, OWASP API Security Top 10, LGPD/GDPR, CI/CD pipeline security, business logic flaws, SAST, secrets history scanning, and infrastructure hardening. Use |
| seo-specialist model: sonnet | SEO specialist covering technical SEO, on-page/off-page optimization, Core Web Vitals, structured data (schema.org), and GEO (SEO for LLMs/AI search). Runs a quality gate whenever the project involves |
| setup-assistant model: opus | Onboards a project into the dev-team-agents ecosystem. Asks the user what type of project it is (new / unfinished / maintenance), configures CLAUDE.md, creates docs/ structure, and optionally integrat |
| software-architect model: opus | Makes architectural decisions after scope is closed. Decides technology stack, system design, patterns, and code standards. Avoids overengineering. Also participates in QUALITY GATE to validate confor |
| technical-writer model: haiku | Produces technical documentation — API docs, READMEs, runbooks, changelogs, and architecture guides. Follows project documentation standards if defined; defaults to Diátaxis framework and Google Devel |
| ui-ux-designer model: sonnet | Dual-role design agent. In Design Mode (pre-build): creates complete visual specifications — design system, component library, typography, color palette, and user flows. In Consultive Mode (alongside |
Hooks (4)
| event | matcher | runs |
|---|---|---|
| PreToolUse | .* | scripts/hooks/pre-tool-use.sh |
| Stop | * | scripts/hooks/stop.sh |
| SessionStart | * | scripts/hooks/session-start.sh |
| PreCompact | * | scripts/hooks/pre-compact.sh |
Slash commands (35)
/adr/architect/audit/backend/commit/dba/design/devops/docs/explain/fix/frontend/fullstack/health-check/install/learn/merge/mobile/plan/pr/push/qa/refactor/relayout/review/rule/security/seo/setup/status/symlinks/sync-rules/tester/update/version
Similar rigs
easonwang00/Startup-Agents
Build a $100K startup with 19 AI employees. Product, Design, Engineering & Marketing agents that share memory and work together. Each agent is a specialized expert with personality, processes, and proven deliverables.
Orchestrator 2.8k tok ·
CronusL-1141/AI-company
Multi-agent team operating system for Claude Code. 108 MCP tools, 40+ agent templates, 10 lifecycle hooks, 7 pipeline workflows. Persistent teams, structured meetings, task wall, real-time React dashboard. No LangChain/AutoGen — pure CC nat
Orchestrator 4.3k tok ·
hoangNguyenAngelhack/ai-agent-setup
🤖 Production-ready Claude Code config — 15 Agents · 17 Rules · 8 Commands · 7 Hooks · 6 Skills
Orchestrator 10.0k tok ·
dominikbullo/ai-setup
My AI setup — Claude Code skills, agents, and config for engineering work
Orchestrator 749 tok ·