Orchestrate AI agents for development tasks using simple slash commands in Claude Code.
Last updated: 2025-08-01
mkdir -p .claude/commands
cp -r /path/to/agentic-dev/.claude/commands/* .claude/commands/
mkdir -p tasks chains/init-project- Analyze and document your project/start-task- Start complex development tasks/complete-task- Finalize and document completed work/test-task- Run adversarial testing
- Commands trigger agent chains (60+ specialized agents)
- Agents research, plan, and execute tasks
- Documentation auto-updates in
tasks/andCLAUDE.md
claude code
> /start-task Add user authentication
# Agents research, plan, and guide implementation
> /complete-task auth "OAuth2 implemented"
# Documentation created, tests optional- Claude Code CLI
- Global
.claude/agents/directory with agents
That's it. Simple, powerful, clean.