msitarzewski/agency-agents

A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.

🎭 The Agency – AI‑Powered Specialist Agents

What it is – A curated collection of markdown files that describe AI agent personalities for a variety of professional roles (frontend developer, AI engineer, security analyst, UI designer, etc.). Each file contains:

  • a short identity and tone of voice,
  • the agent’s core mission and typical workflow,
  • concrete deliverables (code snippets, configuration examples, reports),
  • success‑metrics and communication style.

The repository does not ship runnable code for the agents themselves; instead it provides the prompt‑templates and supporting scripts that let you install those prompts into a number of LLM‑powered coding assistants (Claude Code, Cursor, Gemini CLI, OpenCode, Copilot, etc.).

How you use it

  1. Desktop app (recommended) – The companion app Agency Agents (macOS, Linux, Windows) browses the roster and copies the appropriate markdown files into the configuration folder of the target tool, then keeps them up‑to‑date.
  2. Command‑line installer – ./scripts/install.sh detects which AI‑coding tool you have and copies the selected agents (by division, team, or individual slug) into the tool’s agents/ directory.
  3. Manual copy – For Claude Code you can simply cp engineering/*.md ~/.claude/agents/ and then ask Claude to “activate Frontend Developer mode”.
  4. Conversion – ./scripts/convert.sh generates the tiny JSON/YAML files each supported tool expects.

You can install the whole roster, a whole division (e.g., engineering), or pick individual agents such as frontend-developer or prompt-engineer.

What’s inside

Division Example agents (30‑plus total) Typical use‑case
Engineering Frontend Developer, AI Engineer, DevOps Automator, RAG Pipeline Engineer, Prompt Engineer, Rust Refactoring Specialist, … Generate production‑ready code, design CI/CD pipelines, build retrieval‑augmented generation pipelines, optimise LLM prompts, etc.
Design UI Designer, UX Researcher, Brand Guardian, Whimsy Injector, Image Prompt Engineer, Inclusive Visuals Specialist Produce UI mock‑ups, research user flows, craft brand‑consistent visuals, write image‑generation prompts.
Security, Product, Operations, Marketing, Legal, Finance, HR, Customer Support, Community, Content, Strategy, Management Each contains role‑specific agents (e.g., Incident Response Commander, SEO Specialist, Compliance Analyst, Growth Hacker). Provide domain‑specific guidance, audit artefacts, write policies, run analyses, etc.

All agents are production‑ready in the sense that the author has tested them with the supported LLM tools and documented success metrics.

Who might benefit?

  • Developers who already use an LLM‑powered coding assistant and want ready‑made, role‑specific prompts instead of writing them from scratch.
  • Teams that want a “virtual agency” – a set of specialist personas that can be invoked on demand during a coding session.
  • Product managers looking for a library of AI‑assistant behaviours that can be dropped into internal tooling.

Limitations

  • The repository only supplies prompt definitions (markdown) and thin installer scripts; the actual intelligence lives in the underlying LLM (Claude, Gemini, etc.).
  • Some target tools (e.g., OpenCode) have a hard limit on the number of agents they can register, so you may need to install a subset.
  • No training or fine‑tuning code is provided – it is not a model‑training framework.

Quick start (CLI)

# Install all agents for Claude Code
./scripts/install.sh --tool claude-code

# Install only engineering and security divisions
./scripts/install.sh --tool cursor --division engineering,security

# List available agents
./scripts/install.sh --list agents

License & Community

  • MIT License – free to use, modify, and redistribute.
  • Contributions welcomed (PRs, new agents, bug reports).
  • Sponsorship link and GitHub stars badge shown in the README.

Bottom line – The Agency is a well‑organized, prompt‑library for turning LLM coding assistants into a full‑stack “AI agency” of specialist consultants. It is useful if you already work with Claude Code, Cursor, Gemini CLI, or similar tools and want ready‑made, personality‑rich agents to boost productivity.

Related

  • Project
    addyosmani/agent-skills

    A library of production-grade engineering workflows and quality gates for AI coding agents to ensure they follow senior-level software development best practices.

  • Project
    ajsahni/agents-office

    Agents Office is a beta, locally‑run visual platform that turns Claude into a team of 35 role‑based AI agents. You type tasks, the system routes them to the right agent (or a team), the agents use Claude Code connectors (Gmail, Slack, Chrome, etc.) to do the work, and results are saved as notes in your personal “Brain”. Features include customizable briefs/skills, recurring routines, a calendar view, team‑splitting, model/effort selection, and a non‑commercial license.

  • Project
    MoizIbnYousaf/ai-agent-skills

    A genuine npm‑based CLI/TUI that ships a curated library of AI‑agent skills, lets you build and manage your own skill collections, and integrates with many LLM‑agents via the Agent Skills spec.

  • Project
    msitarzewski/agency-agents-app

    A native desktop installer and manager for AI agent personas that synchronizes specialist personas across various AI coding tools like Cursor and Claude Code.

  • Project
    LambdaTest/agent-skills

    Agent‑Skills is a Node‑JS CLI that adds AI‑assistant‑driven test‑automation “skills” (Selenium, Playwright, Cypress, etc.) for the TestMu AI cloud platform, letting LLMs generate and run tests directly on the service.