Sprint Workflow

An ADDP sprint follows a clear 5-step process. Each phase has clear entry and exit conditions - the orchestrator controls everything automatically.

The sprint in 5 steps

  1. Define tickets
    Each ticket is a markdown file with description, acceptance criteria, and owner files. ADDP can also generate tickets automatically from a concept.
  2. Start sprint
    ADDP creates isolated Git worktrees, sets permissions, and activates the guard hook. The main branch is now protected.
  3. Agents work in parallel
    Each agent works autonomously in its own worktree: tests first, then implementation, build, and commit. No agent can interfere with another.
  4. Automatic review
    Review agents check each worktree for code quality, tests, and acceptance criteria. Result: APPROVED, FIXED, or REJECTED.
  5. Merge in order
    Topological merge into main. Build and tests after each merge. No conflicts thanks to strict isolation.

Wave planning

When tickets have dependencies, waves are formed automatically. Independent tickets run in parallel, dependent tickets start only after completion of the predecessor wave.

Note: The detailed sprint commands, wave planning, and sprint dashboard are described in the included documentation.

Full documentation included with purchase

The detailed configuration reference, troubleshooting guide, and hook documentation are included as PDF in the download.

Buy now →