Skip to content

TeamDay-AI/agents

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TeamDay Agents & Skills Library

Central library for AI agents, skills, and plugins. Agents can be used for:

  • Chat: Direct conversation with specialized AI personas
  • Subagents: Invoked by other agents via Task tool
  • Squads: Pre-composed teams for common use cases

Structure

├── agents/              # 🆕 Agent library
│   ├── registry.json    # Searchable index for discovery
│   ├── schema.ts        # TypeScript types
│   ├── librarian/       # Discovery/onboarding agent
│   ├── specialists/     # Individual expert agents
│   └── squads/          # Pre-composed agent teams
├── skills/              # Skill library (by provider)
│   ├── anthropic/       # git submodule → anthropics/skills
│   ├── huggingface/     # git submodule → huggingface/skills
│   ├── community/       # Community contributed
│   └── teamday/         # Our own skills
└── plugins/             # Claude Code plugins

🤖 Agents

Agents are AI personas with specialized expertise. They can be used as:

  • Chat agents in TeamDay UI (with avatar, greeting, etc.)
  • Subagents invoked via the Task tool in Claude Agent SDK

Available Agents

Agent Category Description
📚 Agent Librarian productivity Discovers and installs agents/squads
🔍 SEO Specialist marketing Search performance and optimization
✍️ Content Writer marketing Blog posts, copy, documentation
🎨 Frontend Developer development UI/UX with React, Vue, Tailwind
⚙️ Backend Developer development APIs, databases, security
🚀 DevOps Engineer operations CI/CD, cloud, infrastructure
📊 Data Analyst data Analytics, SQL, visualization

Squads (Agent Teams)

Squad Agents Use Case
🏆 SEO Website Squad SEO + Content + Frontend Build SEO-optimized websites
🔧 Full-Stack Squad Frontend + Backend + DevOps Build complete web applications

Agent Format

Agents are defined as Markdown files with YAML frontmatter:

---
id: seo-specialist
name: SEO Specialist
description: Analyzes search performance...  # For auto-invocation
version: 1.0.0
avatar: "🔍"                                  # Emoji or image URL
greeting: |                                   # First message in chat
  Hey! I'm your SEO Specialist...
category: marketing
tags: [seo, analytics, keywords]
tier: pro                                     # free | pro | enterprise
tools: [Read, Glob, Grep, WebSearch]          # Allowed tools
model: sonnet                                 # sonnet | opus | haiku
requires:
  mcps: [search-console]                      # Required MCP servers
  credentials: [GOOGLE_ANALYTICS_CREDENTIALS] # Required secrets
worksWellWith: [content-writer, data-analyst] # Complementary agents
---

# SEO Specialist

[System prompt defining the agent's behavior...]

Using Agents

In TeamDay (Chat): Agents appear in the agent selector with their avatar and greeting.

As Subagent (SDK):

import { query } from '@anthropic-ai/claude-agent-sdk'

for await (const msg of query({
  prompt: "Use the seo-specialist agent to audit this site",
  options: {
    agents: {
      'seo-specialist': {
        description: 'SEO analysis and optimization',
        prompt: '...', // Loaded from .md file
        tools: ['Read', 'Grep', 'WebSearch']
      }
    }
  }
})) { ... }

Install to Space:

cp agents/specialists/seo-specialist.md {space}/.claude/agents/

Skills by Provider

anthropic (git submodule)

Official Anthropic skills from anthropics/skills:

Skill Description
anthropic/pdf PDF processing, form filling, merge/split
anthropic/docx Word document creation
anthropic/xlsx Excel spreadsheet operations
anthropic/pptx PowerPoint presentations
anthropic/mcp-builder Build MCP servers
anthropic/webapp-testing Web automation testing

To update: cd skills/anthropic && git pull

huggingface (git submodule)

Official HuggingFace skills from huggingface/skills:

Skill Description
huggingface/hugging-face-datasets Create and manage HF datasets
huggingface/hugging-face-model-trainer Train models on HF
huggingface/hugging-face-evaluation Evaluate models
huggingface/hugging-face-paper-publisher Publish papers to HF
huggingface/hugging-face-tool-builder Build HF tools

To update: cd skills/huggingface && git pull

teamday

TeamDay custom skills:

Skill Description
teamday/skills-manager Import and maintain skills from various sources

Skill Naming Convention

Skills are referenced as {provider}/{skill-name}:

  • anthropic/pdf
  • anthropic/docx
  • teamday/seo-analyst

Or with prefix for flat namespaces:

  • anthropic-pdf
  • teamday-seo-analyst

Usage

For TeamDay App

When installing a skill to a space:

# Copy skill directory
cp -r skills/anthropic/skills/pdf /sandbox/{org}/{space}/.claude/skills/pdf

Agent references:

skills: ["pdf", "docx"]  # or ["anthropic/pdf", "anthropic/docx"]

For Claude Code

/plugin marketplace add TeamDay-AI/agents
/plugin install blog-image-generator@teamday-agents

Adding New Providers

# Add as submodule
git submodule add https://github.com/{org}/{repo}.git skills/{provider-name}

Updating Skills

# Update all submodules
git submodule update --remote

# Update specific provider
cd skills/anthropic && git pull origin main

Cloning This Repo

# Clone with submodules
git clone --recursive https://github.com/TeamDay-AI/agents.git

# Or init submodules after clone
git clone https://github.com/TeamDay-AI/agents.git
cd agents
git submodule init
git submodule update

About

AI agents

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors