From dream-team
Autonomous sprint execution with checkpoint-based progress tracking
npx claudepluginhub numanumanuma/claude-plugins --plugin dream-teamThis skill uses the workspace's default tool permissions.
The sprint runner (`scripts/run-task.sh`) enables autonomous execution of sprint tasks. It reads the checkpoint from `progress.md`, launches Claude with fresh context each iteration, and loops until the sprint is complete or blocked.
Verifies tests pass on completed feature branch, presents options to merge locally, create GitHub PR, keep as-is or discard; executes choice and cleans up worktree.
Guides root cause investigation for bugs, test failures, unexpected behavior, performance issues, and build failures before proposing fixes.
Writes implementation plans from specs for multi-step tasks, mapping files and breaking into TDD bite-sized steps before coding.
The sprint runner (scripts/run-task.sh) enables autonomous execution of sprint tasks. It reads the checkpoint from progress.md, launches Claude with fresh context each iteration, and loops until the sprint is complete or blocked.
<!-- CHECKPOINT --> block from progress.md to determine current stateprogress.md (checkpoint + log) and tasks.md (checkboxes)phase: done or phase: blockedSPRINT_DIR=planning/sprints/sprint-N-name
mkdir -p $SPRINT_DIR
Copy the templates from this plugin's templates/ directory:
plan.md — Architecture decisions, phases, riskstasks.md — Task checklist with status trackingprogress.md — Checkpoint block + progress logComplete plan.md with architecture decisions and phases. Fill in tasks.md with all implementation tasks. Initialize progress.md with the first checkpoint.
./scripts/run-task.sh planning/sprints/sprint-N-name
The checkpoint is the ONLY state that persists between runner iterations. Make it precise:
<!-- CHECKPOINT
sprint: 5
sprint_name: dark-mode
active_task: "Implement theme provider service"
phase: implementing
last_completed: "Created ThemeConfig model with color palette and font definitions"
next_step: "Create ThemeProvider service with system preference detection and manual toggle"
blockers: none
files_modified: src/models/ThemeConfig.ts
-->
active_task: Exact task name from tasks.mdnext_step: Specific enough that a fresh session can start immediatelyfiles_modified: Every file touched in the current session (helps the next iteration know what to read)The runner handles bug fixes with a TDD workflow:
next_step in the checkpoint points to the next task, not a vague description