This repository contains shared Claude Code workflow templates:

• project settings
• commands
• subagents
• hooks
• MCP examples

It does not ship a backend service or private hosted component.

If you find a security issue in this starter, open a private report through the repository security workflow if enabled, or open a normal issue with enough detail to reproduce the problem safely.

Please report:

• hooks that can leak secrets or execute unsafe commands by default
• examples that encourage committing credentials
• unsafe permission defaults
• misleading guidance that could cause destructive behavior

• no committed secrets
• no default network exfiltration hooks
• no destructive shell behavior without an explicit ask/deny flow
• no hidden telemetry

This starter is only a baseline.

Each team should review:

• .claude/settings.json
• .mcp.json
• environment variables used by MCP
• any custom hooks added after adoption