
agency-agents-ai-specialists
A collection of specialized AI agent personalities for Claude Code, Cursor, Aider, Windsurf, and other AI coding tools — covering engineering, design, marketing, sales, and more.
- Installs
- 2.1K
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/reason-machines/trending-skills --skill agency-agents-ai-specialists- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads agency-agents-ai-specialists automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
🎭 Agency Agents — AI Specialist Personalities
Skill by ara.so — Daily 2026 Skills collection.
A curated collection of 50+ specialized AI agent personalities for Claude Code, Cursor, Aider, Windsurf, Copilot, and more. Each agent has deep domain expertise, a distinct personality, defined workflows, and measurable deliverables — covering engineering, design, marketing, sales, paid media, and beyond.
Installation
Prerequisites
git clone https://github.com/msitarzewski/agency-agents.git
cd agency-agents
Claude Code (Recommended)
# Copy all agents to Claude's agents directory
cp -r agency-agents/* ~/.claude/agents/
# Or symlink for auto-updates
ln -s /path/to/agency-agents ~/.claude/agents/agency
Then in any Claude Code session:
Hey Claude, activate Frontend Developer mode and help me build a React component
All Other Tools (Interactive Installer)
# Step 1: Generate integration files for all supported tools
./scripts/convert.sh
# Step 2: Auto-detect installed tools and install interactively
./scripts/install.sh
# Or target a specific tool
./scripts/install.sh --tool cursor
./scripts/install.sh --tool copilot
./scripts/install.sh --tool aider
./scripts/install.sh --tool windsurf
Manual per Tool
| Tool | Install path |
|---|---|
| Claude Code | ~/.claude/agents/ |
| Cursor | .cursor/rules/ in project root |
| Copilot | .github/copilot-instructions.md |
| Aider | .aider.conf.yml or pass via --system-prompt |
| Windsurf | .windsurf/rules/ in project root |
Agent Roster
Engineering Division
engineering/engineering-frontend-developer.md React/Vue/Angular, UI, Core Web Vitals
engineering/engineering-backend-architect.md API design, databases, scalability
engineering/engineering-mobile-app-builder.md iOS/Android, React Native, Flutter
engineering/engineering-ai-engineer.md ML models, AI integration, data pipelines
engineering/engineering-devops-automator.md CI/CD, infra automation, cloud ops
engineering/engineering-rapid-prototyper.md MVPs, POCs, hackathon speed
engineering/engineering-senior-developer.md Laravel/Livewire, advanced patterns
engineering/engineering-security-engineer.md Threat modeling, secure code review
engineering/engineering-code-reviewer.md PR reviews, code quality gates
engineering/engineering-database-optimizer.md PostgreSQL/MySQL tuning, slow queries
engineering/engineering-git-workflow-master.md Branching, conventional commits
engineering/engineering-software-architect.md System design, DDD, trade-off analysis
engineering/engineering-sre.md SLOs, error budgets, chaos engineering
engineering/engineering-incident-response-commander.md Incident management, post-mortems
engineering/engineering-technical-writer.md Developer docs, API reference
engineering/engineering-data-engineer.md Data pipelines, lakehouse, ETL/ELT
Design Division
design/design-ui-designer.md Visual design, component libraries
design/design-ux-researcher.md User testing, behavior analysis
design/design-ux-architect.md CSS systems, technical UX
design/design-brand-guardian.md Brand identity and consistency
design/design-whimsy-injector.md Micro-interactions, delight, Easter eggs
design/design-image-prompt-engineer.md Midjourney/DALL-E/SD prompts
design/design-inclusive-visuals-specialist.md Representation, bias mitigation
Marketing, Sales & Paid Media
marketing/marketing-growth-hacker.md
marketing/marketing-content-creator.md
paid-media/paid-media-ppc-strategist.md
paid-media/paid-media-creative-strategist.md
sales/sales-outbound-strategist.md
sales/sales-deal-strategist.md
sales/sales-discovery-coach.md
Using Agents in Claude Code
Activating a Single Agent
# In Claude Code chat:
Activate the Backend Architect agent and help me design a REST API for a multi-tenant SaaS app.
Using Multiple Agents in Sequence
# First, design the system
Activate the Software Architect agent. Design the domain model for an e-commerce platform.
# Then implement
Now activate the Senior Developer agent and implement the Order aggregate in Laravel.
# Then review
Activate the Code Reviewer agent and review the implementation above.
Referencing an Agent File Directly
# Pass an agent as a system prompt in Claude CLI
claude --system-prompt "$(cat ~/.claude/agents/engineering-frontend-developer.md)" \
"Build a responsive product card component in React with Tailwind CSS"
Using Agents in Cursor
After running ./scripts/install.sh --tool cursor, agent rules land in .cursor/rules/. Reference them in chat:
@engineering-frontend-developer Build a data table component with sorting and pagination.
Or set a default rule in .cursor/rules/default.mdc:
---
alwaysApply: true
---
You are operating as the Senior Developer agent from The Agency.
Refer to .cursor/rules/engineering-senior-developer.md for your full persona and workflows.
Using Agents with Aider
# Use a single agent as the system prompt
aider --system-prompt "$(cat agency-agents/engineering/engineering-security-engineer.md)"
# Or reference in .aider.conf.yml
echo "system-prompt: agency-agents/engineering/engineering-devops-automator.md" >> .aider.conf.yml
Using Agents in Windsurf
./scripts/install.sh --tool windsurf
# Agents are written to .windsurf/rules/
Activate in chat:
Use the UX Architect agent rules from .windsurf/rules/ to audit my CSS architecture.
Real Workflow Examples
Full-Stack Feature with Multiple Agents
# 1. Architecture phase
cat > task.md << 'EOF'
I need to add real-time notifications to my Node.js + React app.
Users should see in-app alerts and optionally receive email digests.
EOF
# Invoke Software Architect
claude --system-prompt "$(cat ~/.claude/agents/engineering-software-architect.md)" < task.md
# 2. Backend implementation
claude --system-prompt "$(cat ~/.claude/agents/engineering-backend-architect.md)" \
"Implement the notification service based on the architecture above using PostgreSQL LISTEN/NOTIFY and Socket.io"
# 3. Frontend implementation
claude --system-prompt "$(cat ~/.claude/agents/engineering-frontend-developer.md)" \
"Build the React notification bell component that connects to the Socket.io feed"
# 4. Security review
claude --system-prompt "$(cat ~/.claude/agents/engineering-security-engineer.md)" \
"Review the notification system implementation for security issues"
Code Review Workflow
# Generate a diff and pipe to the Code Reviewer agent
git diff main..feature/payment-integration | \
claude --system-prompt "$(cat ~/.claude/agents/engineering-code-reviewer.md)" \
"Review this PR diff. Focus on security, correctness, and maintainability."
Database Optimization
# Paste slow query log and activate Database Optimizer
claude --system-prompt "$(cat ~/.claude/agents/engineering-database-optimizer.md)" << 'EOF'
Here is a slow query from our PostgreSQL logs (avg 4200ms):
SELECT u.*, p.*, o.*
FROM users u
LEFT JOIN profiles p ON p.user_id = u.id
LEFT JOIN orders o ON o.user_id = u.id
WHERE u.created_at > NOW() - INTERVAL '30 days'
ORDER BY o.created_at DESC;
Table sizes: users=2M rows, orders=18M rows. No indexes on created_at columns.
EOF
Incident Response
# Structured incident kick-off
claude --system-prompt "$(cat ~/.claude/agents/engineering-incident-response-commander.md)" << 'EOF'
SEV-1 INCIDENT: Payment processing returning 503 errors since 14:32 UTC.
Error rate: 94%. Affected: checkout, subscription renewals.
Recent deploys: payment-service v2.4.1 at 14:15 UTC.
EOF
Creating Custom Agents
Agent files follow a consistent markdown structure:
# 🎯 Agent Name
## Identity
You are [Name], [role] at The Agency...
## Core Mission
[What this agent optimizes for]
## Personality & Communication Style
- [Trait 1]
- [Trait 2]
## Workflows
### [Workflow Name]
1. [Step 1]
2. [Step 2]
## Deliverables
- [Concrete output 1]
- [Concrete output 2]
## Success Metrics
- [Measurable outcome]
Save custom agents to agency-agents/custom/ and re-run ./scripts/convert.sh to generate tool integrations.
Contributing New Agents
# Fork and clone
git clone https://github.com/YOUR_USERNAME/agency-agents.git
# Create your agent in the appropriate division
cp engineering/engineering-senior-developer.md \
engineering/engineering-YOUR-SPECIALTY.md
# Edit the file, then test it
claude --system-prompt "$(cat engineering/engineering-YOUR-SPECIALTY.md)" \
"Give me a sample deliverable to demonstrate your capabilities"
# Submit a PR
git checkout -b agent/your-specialty
git add engineering/engineering-YOUR-SPECIALTY.md
git commit -m "feat: add Your Specialty agent"
git push origin agent/your-specialty
Troubleshooting
Agents not found in Claude Code
ls ~/.claude/agents/
# If empty, re-run:
cp -r /path/to/agency-agents/* ~/.claude/agents/
convert.sh fails with permission error
chmod +x scripts/convert.sh scripts/install.sh
./scripts/convert.sh
Cursor not picking up agent rules
# Rules must be in project root .cursor/rules/
ls .cursor/rules/
# Re-run installer targeting cursor
./scripts/install.sh --tool cursor
Agent personality not activating
- Be explicit: "Activate the Frontend Developer agent" rather than just referencing the topic
- Paste the agent file contents directly into the system prompt if tool integration isn't working
- For Claude Code, confirm agents directory:
claude config get agentsDir
Agent conflicts when using multiple
- Activate one agent per conversation session
- For multi-agent workflows, use separate sessions or Claude Code's subagent feature
- Sequence agents explicitly: architect → implement → review
Project Structure
agency-agents/
├── engineering/ # 23 engineering specialist agents
├── design/ # 8 design specialist agents
├── marketing/ # Marketing and growth agents
├── sales/ # 8 sales specialist agents
├── paid-media/ # 7 paid media specialist agents
├── scripts/
│ ├── convert.sh # Generate tool-specific integration files
│ └── install.sh # Interactive installer (auto-detects tools)
└── README.md
Key Facts
- License: MIT
- 51,000+ stars — battle-tested by a large community
- No API keys required — agents are prompt files, not services
- Tool-agnostic — works with any LLM tool that accepts system prompts
- Extensible — add custom agents following the same markdown pattern
- PRs welcome — the roster grows through community contributions
FAQ
What does the agency-agents-ai-specialists skill do?
A collection of specialized AI agent personalities for Claude Code, Cursor, Aider, Windsurf, and other AI coding tools — covering engineering, design, marketing, sales, and more.
How do I install agency-agents-ai-specialists?
Run npx skills add https://github.com/reason-machines/trending-skills --skill agency-agents-ai-specialists in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/agency-agents-ai-specialists/ (personal) or .claude/skills/agency-agents-ai-specialists/ (project).
Which AI agents can use agency-agents-ai-specialists?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is agency-agents-ai-specialists safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains agency-agents-ai-specialists?
It is published by reason-machines in reason-machines/trending-skills. It has 2,112 recorded installs.
More from reason-machines/trending-skills
openclaw-control-center
reason-machines/trending-skills
Local-first, security-first control center for OpenClaw agents — visibility dashboard with readonly defaults, token attribution, collaboration tracing, and safe write operations.
inkos-multi-agent-novel-writing
reason-machines/trending-skills
Multi-agent CLI system for autonomous novel writing, auditing, and revision with human review gates
claude-hud-statusline
reason-machines/trending-skills
A Claude Code plugin that displays a real-time HUD showing context usage, active tools, running agents, and todo progress in your terminal statusline.
understand-anything-knowledge-graph
reason-machines/trending-skills
Turn any codebase into an interactive knowledge graph using Claude Code skills — explore, search, and ask questions about any project visually.
agent-browser-automation
reason-machines/trending-skills
Headless browser automation CLI for AI agents using native Rust binary with Chrome DevTools Protocol
ui-ux-pro-max-skill
reason-machines/trending-skills
AI design intelligence skill for building professional UI/UX across multiple platforms with 161 reasoning rules, 67 styles, and automated design system generation
Popular ai & agents skills
google-agents-cli-publish
google/agents-cli
This skill should be used when the user wants to "publish an agent", "publish my ADK agent", "register an agent with Gemini", "publish to Gemini" (formerly Gemini Enterprise), or needs guidance on the agents-cli publish gemini command. Also use when the user wants to "manage agents in Agent Registry", "list/update/delete registered agents", or "register an MCP server". Covers ADK vs A2A registration modes, programmatic and interactive usage, flag reference, auto-detection from deployment metadata, Agent Registry fleet management, and troubleshooting. Part of the agents-cli skills suite. Do NOT use for deployment (use google-agents-cli-deploy).
google-agents-cli-eval
google/agents-cli
This skill should be used when the user wants to "run an evaluation", "evaluate my agent", "evaluate my ADK agent", "write an eval dataset", "analyze eval failures", "compare eval results", "optimize agent", or needs guidance on the Gemini platform eval methodology and the Quality Flywheel. Covers eval metrics, dataset schema, LLM-as-judge scoring, and common failure causes. Applies to any agents-cli project, whatever framework the agent is written in. Do NOT use for agent API code patterns (ADK: use google-agents-cli-adk-code), deployment (use google-agents-cli-deploy), or project scaffolding (use google-agents-cli-scaffold).
skill-creator
anthropics/skills
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
codex-pet
prime-skills/runcomfy-agent-skills
Codex Pet generator on RunComfy. Build a Codex-compatible Codex Pet spritesheet.webp + pet.json from a single reference image, drop it into ${CODEXHOME:-$HOME/.codex}/pets/ / and Codex picks it up as a custom Codex Pet next to the 8 built-ins. This skill produces the exact Codex Pet atlas Codex expects (1536x1872 PNG/WebP, 8 cols x 9 rows, 192x208 cells, 9 animation states — idle, running-right, running-left, waving, jumping, failed, waiting, running, review). Calls OpenAI GPT Image 2 edit ONCE via the local RunComfy CLI as runcomfy run openai/gpt-image-2/edit to produce a canonical Codex Pet pose, then assembles all 9 animation rows programmatically with ImageMagick micro-transforms — no Codex Pro, no $imagegen, no OPENAIAPIKEY required, only RUNCOMFYTOKEN. Triggers on "codex pet", "create codex pet", "make codex pet", "hatch codex pet", "/hatch image", "desktop pet codex", "codex pets", "spritesheet.webp", or any explicit ask to build a custom pet for OpenAI Codex.
full-output-enforcement
leonxlnx/taste-skill
Overrides default LLM truncation behavior. Enforces complete code generation, bans placeholder patterns, and handles token-limit splits cleanly. Apply to any task requiring exhaustive, unabridged output.
executing-plans
obra/superpowers
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available