Claude Code Starter Kit v1.1.0

v1.1.0 — MCP Server

This release makes the Claude Code Starter Kit pip-installable as an MCP server, so all commands and agents can be used from Claude Desktop, Claude Code, or any
MCP-compatible client — without cloning the repo.

Install

pip install claude-code-starter-kit-mcp

What's included

8 command tools — each maps to a slash command from the kit:
command_test, command_test_coverage, command_security_check, command_detailed_review, command_fix_bugs, command_refactor, command_docs, command_monitor

4 agent tools — specialized AI personas:
agent_test_engineer, agent_security_auditor, agent_code_reviewer, agent_bug_hunter

Setup

Add to your Claude config (~/.claude/settings.json for Claude Code, or claude_desktop_config.json for Claude Desktop):

{
"mcpServers": {
"claude-code-starter-kit": {
"command": "claude-code-starter-kit-mcp"
}
}
}

Changes

• Added mcp_server.py — stdio MCP server that dynamically loads commands and agents from bundled markdown files
• Added pyproject.toml — package metadata, entry point, and wheel configuration
• Added GitHub Actions workflow for automated PyPI publishing on tag push

Claude Code Starter Kit v1.0.0

🚀 v1.0.0 - Initial Release

Commands (8)

• /test - Generate unit tests
• /test-coverage - Improve coverage
• /detailed-review - Structured code review
• /security-check - Dependency vulnerabilities
• /fix-bugs - Auto-fix issues
• /docs - Generate documentation
• /refactor - Safe refactoring
• /monitor - Repo health check

Agents (4)

• test-engineer
• security-auditor
• code-reviewer
• bug-hunter

💡 For quick reviews, use Claude Code's built-in /review and /security-review Opus 4.5Extended