Why ADDP?

AI agents are powerful - but without structure they create chaos. Anyone who has let multiple agents loose on a repo at the same time knows the problems.

The pain points

• Agent collisions: Two agents modify the same file. Merge conflicts. One overwrites the other. Debugging instead of progress.
• Missing quality control: Agents write code without tests. No review. Bugs slip in and cost ten times more later.
• Sequential instead of parallel: One agent at a time. With 10 tickets it takes 10x as long. Parallelism fails without isolation.
• Broken commits on main: Agents commit code that does not build. Merge failures on main block the entire team.

With vs. without ADDP

Aspect	Without ADDP	With ADDP
Isolation	All agents on main	Each agent in its own worktree
Tests	Optional, often forgotten	TDD gate enforces tests before code
Review	Manual or none	Automatic review agent
Parallelism	Conflicts from agent 2	Unlimited parallel, conflict-free
Build quality	Broken builds possible	BuildCheck blocks red commits
Setup	Custom scripts needed	One command: /addp

What ADDP includes

• 8 safety hooks: Guard, TDD Gate, BuildCheck, SkipTest Detector, NoStash, MergeVerify, TeammateIdle, TaskCompleted
• 30+ languages: TDD Gate and SkipTest Detector recognize test patterns in Go, TypeScript, Python, Rust, C++, Java, C#, Ruby, Swift, PHP, Dart and more
• 6 platforms: macOS, Linux, Windows - ARM64 and AMD64 each
• 2 modes: Agent Teams (real parallel sessions) or Subagent Fallback
• Signed installers: macOS .pkg (notarized), Windows .bat, Linux .sh

Who is ADDP for?

• Developers using Claude Code for larger projects
• Teams that need parallel agent workflows
• Projects where code quality is not optional