This directory contains examples of how to integrate bd with AI agents and workflows.
- python-agent/ - Simple Python agent that discovers ready work and completes tasks
- bash-agent/ - Bash script showing the full agent workflow
- startup-hooks/ - Session startup scripts for automatic bd upgrade detection
- claude-desktop-mcp/ - MCP server for Claude Desktop integration
- monitor-webui/ - Standalone web interface for real-time issue monitoring and visualization
- git-hooks/ - Pre-configured git hooks for automatic Dolt sync
- contributor-workflow/ - OSS contributor setup with separate planning repo
- team-workflow/ - Team collaboration with shared repositories
- multi-phase-development/ - Organize large projects by phases (planning, MVP, iteration, polish)
- multiple-personas/ - Architect/implementer/reviewer role separation
- protected-branch/ - Protected branch workflow for team collaboration
# Try the Python agent example
cd python-agent
python agent.py
# Try the bash agent example
cd bash-agent
./agent.sh
# Install git hooks
cd git-hooks
./install.sh
# REMOVED (bd-4c74): branch-merge demo - hash IDs eliminate collision resolutionThe basic agent workflow:
- Find ready work:
bd ready --json --limit 1 - Claim the task:
bd update <id> --claim --json - Do the work: Execute the task
- Discover new issues:
bd create "Found bug" --json - Link discoveries:
bd dep add <new-id> <parent-id> --type discovered-from - Complete the task:
bd close <id> --reason "Done" --json
All commands support --json for easy parsing.