Files
Daniel Avila ebf6c519b7 improve: enhance bettoredge-value-finder agent component (#728)
* improve: enhance bettoredge-value-finder based on automated review

- Fix name to kebab-case (bettoredge-value-finder) matching filename
- Add required tools/model frontmatter, scoped to Read + mcp__bettoredge__*
- Add system-prompt persona and "When invoked" workflow
- Add Guardrails section enforcing bankroll limits, no fabricated results,
  and legal-age confirmation
- Caveat the bid/ask-midpoint value-detection methodology
- Add credential-handling guidance for live account login env vars
- Bump version to 1.0.1 to match published npm package

Automated review cycle | Co-Authored-By: Claude Code <noreply@anthropic.com>

* fix: address PR review feedback on bettoredge-value-finder

- Remove unnecessary Read from tools frontmatter (agent only calls
  MCP tools, none of which touch the filesystem)
- Fix bootstrap paradox: step 1 no longer unconditionally calls
  bettoredge_setup, which only exists once the MCP server is running
- Move credential-handling warning ahead of the first instruction
  that has users embed BETTOREDGE_EMAIL/PASSWORD

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: clarify bettoredge_status is read-only in guardrail wording

bettoredge_status only surfaces current bankroll limits; it can't
change them. Point users to the BettorEdge platform for that instead
of implying the agent can modify limits through this tool.

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-07-16 10:07:42 -04:00
..
2026-07-14 19:25:34 -04:00
…
…
2026-07-14 19:25:34 -04:00
…

npm version npm downloads License: MIT Open Source PRs Welcome GitHub stars

Claude Code Templates

CLI tool for configuring and monitoring Claude Code - Quick setup for any project with framework-specific commands and real-time monitoring dashboard.

🚀 Quick Start

# Interactive setup (recommended)
npx claude-code-templates@latest

# Real-time analytics dashboard
npx claude-code-templates@latest --analytics

# System health check
npx claude-code-templates@latest --health-check

✨ Core Features

  • 📋 Smart Project Setup - Auto-detect and configure any project with framework-specific commands
  • 📊 Real-time Analytics - Monitor Claude Code sessions with live state detection and performance metrics
  • 🔍 Health Check - Comprehensive system validation with actionable recommendations
  • 🧩 Individual Components - Install specialized agents, commands, and MCPs individually
  • 🌍 Global Agents - Create AI agents accessible from anywhere using Claude Code SDK

🎯 What You Get

Component Description Example
CLAUDE.md Project-specific Claude Code configuration Framework best practices, coding standards
Commands Custom slash commands for development tasks /generate-tests, /check-file, /optimize-bundle
Agents AI specialists for specific domains API security audit, React performance, database optimization
MCPs External service integrations GitHub, databases, development tools
Skills Modular capabilities with progressive disclosure PDF processing, algorithmic art, MCP builder
Analytics Real-time monitoring dashboard Live session tracking, usage statistics, exports

🛠️ Supported Technologies

Language Frameworks Status
JavaScript/TypeScript React, Vue, Angular, Node.js ✅ Ready
Python Django, Flask, FastAPI ✅ Ready
Common Universal configurations ✅ Ready
Go Gin, Echo, Fiber 🚧 Coming Soon
Rust Axum, Warp, Actix 🚧 Coming Soon

🌍 Global Agents (Claude Code SDK Integration)

Create AI agents that can be executed from anywhere using the Claude Code SDK:

# Create a global agent (one-time setup)
npx claude-code-templates@latest --create-agent customer-support

# Use the agent from anywhere
customer-support "Help me with ticket #12345"
sre-logs "Analyze error patterns in app.log"  
code-reviewer "Review this PR for security issues"

Available Global Agents

Agent Usage Description
customer-support customer-support "query" AI customer support specialist
api-security-audit api-security-audit "analyze endpoints" Security auditing for APIs
react-performance-optimization react-performance-optimization "optimize components" React performance expert
database-optimization database-optimization "improve queries" Database performance tuning

Global Agent Management

# List installed global agents
npx claude-code-templates@latest --list-agents

# Update an agent to latest version
npx claude-code-templates@latest --update-agent customer-support

# Remove an agent
npx claude-code-templates@latest --remove-agent customer-support

How It Works

  1. Download Agent: Fetches the latest agent from GitHub
  2. Generate Executable: Creates a Node.js script that calls Claude Code SDK
  3. Add to PATH: Makes the agent available globally in your shell
  4. Ready to Use: Execute agent-name "your prompt" from any directory

The agents use the Claude Code SDK internally to provide specialized AI assistance with domain-specific knowledge and best practices.

🎨 Skills (Anthropic Format)

Install modular capabilities that Claude loads dynamically using Anthropic's progressive disclosure pattern:

# Install individual skills
npx claude-code-templates@latest --skill pdf-processing-pro
npx claude-code-templates@latest --skill algorithmic-art
npx claude-code-templates@latest --skill mcp-builder

# Install multiple skills
npx claude-code-templates@latest --skill pdf-anthropic,docx,xlsx,pptx

🎨 Creative & Design

  • algorithmic-art - Create generative art using p5.js with seeded randomness
  • canvas-design - Design beautiful visual art in .png and .pdf formats
  • slack-gif-creator - Create animated GIFs optimized for Slack

💻 Development & Technical

  • mcp-builder - Guide for creating high-quality MCP servers
  • artifacts-builder - Build complex HTML artifacts with React and Tailwind
  • webapp-testing - Test local web applications using Playwright
  • skill-creator - Guide for creating effective skills

📄 Document Processing

  • pdf-processing-pro - Production-ready PDF toolkit (forms, tables, OCR)
  • pdf-anthropic - Anthropic's comprehensive PDF manipulation toolkit
  • docx - Create, edit, and analyze Word documents
  • xlsx - Create, edit, and analyze Excel spreadsheets
  • pptx - Create, edit, and analyze PowerPoint presentations

🏢 Enterprise & Communication

  • brand-guidelines - Apply Anthropic's official brand guidelines
  • internal-comms - Write internal communications (reports, newsletters, FAQs)
  • theme-factory - Style artifacts with professional themes

Skills Architecture

Skills follow Anthropic's progressive disclosure pattern:

  • Metadata - Always loaded (name, description)
  • Instructions - Loaded when skill is triggered
  • Resources - Reference files loaded only when needed
  • Scripts - Execute without loading code into context

Attribution

Skills from anthropics/skills:

  • Open Source (Apache 2.0): algorithmic-art, mcp-builder, skill-creator, artifacts-builder, and more
  • Source-Available (Reference): docx, pdf-anthropic, pptx, xlsx

See ANTHROPIC_ATTRIBUTION.md for complete license information.

📖 Documentation

📚 Complete Documentation - Comprehensive guides, examples, and API reference

Quick links:

🤝 Contributing

We welcome contributions! Browse available templates and components at aitmpl.com, then check our contributing guidelines.

📄 License

MIT License - see the LICENSE file for details.


⭐ Found this useful? Give us a star to support the project!