By mthanhlm
Scrum + XP workflow: sprints and a backlog for management; TDD, codegraph-grounded stories, a navigator review pass, and an enforced done-gate for quality.
Finish the active story — navigator review, then the done-gate (verify set). Closes only when both pass.
Initialise ultrapower in this project — detect verify commands and the Definition of Done, scaffold .scrum/.
Set the lean ladder's intensity (lite/full/ultra/off). Always-on; the level persists per project.
Refactor safely — codegraph impact first, lock scope, restructure with tests green throughout. No behavior change.
Run a short sprint retrospective and append the notes to .scrum/retro.md.
Read-only pre-lock critic. Reviews the plan, not the diff — six distinct lenses across a story's BRIEF and proposed file contract before anything is locked or coded. Returns severity-tagged findings for the human to resolve. Never edits code.
Executes ONE locked story test-first — red (failing test) → green (minimal code) → refactor. Stays inside the story's file contract, runs the verify commands, and reports back. Use after a story brief is approved and locked.
Read-only reviewer (the XP pairing "navigator"). Reviews a finished story's diff against its brief and contract before the done-gate — six lenses applied to the FINISHED DIFF — and returns severity-tagged findings. Never edits code.
Facilitates sprint planning and close. Turns intake answers into a crisp sprint goal and a clear, estimated backlog; computes velocity on close. Returns a draft for the main loop to confirm — never writes files, never addresses the user directly.
Read-only planner for a single story. Grounds in codegraph (impact + reuse), produces a tight brief (what/deliverable/files/affected sites/out-of-scope/verify) and a proposed point estimate. Never edits code.
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that fuses Scrum (lightweight management) with Extreme Programming (quality engineering) into one enforced workflow: plan a sprint, drive each story test-first with codegraph grounding, get a navigator review, and pass a verify-gate before anything closes — all under an always-on lean ladder (YAGNI → stdlib → native → one line) that keeps every change minimal.
Commands are namespaced /up: (the plugin is named up).
/up:init once per project — detect verify commands + Definition of Done
/up:lean [level] dial the always-on lean ladder (lite|full|ultra|off; default full)
/up:sprint plan guided question-bundle → crisp sprint goal + estimated backlog
/up:story start <id> brief + codegraph impact/reuse → you approve → debate pre-lock plan review → file contract locked
(implement) TDD: failing test → mark-red → minimal code → refactor
/up:done navigator + done-gate → auto-mark done, release lock, queue for tutoring
/up:sprint close record velocity → /up:retro
/up:tutor [target] understand the code to mastery — or drain the pending queue → tutored.md
New to it? Here's a full pass for a small task: "reject malformed emails on the signup form."
0. One-time setup (per project)
/up:init
It scans the repo, proposes the commands it found (e.g. pytest -q, ruff check ., mypy, and a
run/boot smoke), and asks you to confirm. Writes .scrum/config.json.
1. Plan the sprint
/up:sprint plan
It interviews you with a few quick questions (what's the goal? what's in/out of scope? how big?),
then proposes a sprint goal and a short backlog with point estimates — e.g.
"S1 — reject malformed emails on signup (2 pts)". You tweak and approve; it writes .scrum/sprint.md.
2. Start a story
/up:story start S1
The story-planner reads the codebase via codegraph and hands you a brief: which files it'll
touch, the approach, which existing helpers it reuses, the tests, and a verify command. You
approve → the debate agent runs a pre-lock plan review (six lenses, VERDICT: proceed | revise-plan) → that file list is locked (edits outside it get blocked).
3. Build it test-first
test_rejects_malformed_email → run it → it fails (red). Required.If you try to edit the validator before a failing test exists, the tdd-guard stops you — that's the plugin keeping you honest, not a bug. (No active story? All guards are off.)
4. Finish the story
/up:done
The navigator reviews your diff against the brief (missing test? a caller you forgot? scope
creep?) and lists findings. Fix any blockers, then the done-gate runs your verify set
(tests + lint + typecheck + smoke). Green + no blockers → the story is auto-marked done, the
lock releases, the increment is tagged, and the story is queued for tutoring — work the queue
anytime with /up:tutor. No hand-editing of .scrum/ state.
5. Repeat & wrap up
/up:sprint close # records velocity for next sprint's planning
/up:retro # quick 3-bullet retrospective
Reach for these as needed:
/up:refactor <target> — restructure with no behavior change (checks every caller via codegraph first, keeps tests green)./up:story add "<idea>" — capture a backlog idea without planning it yet./up:tutor [target] — a tutor drills you to mastery over a story's diff, a file/area, the whole project, or a topic; with no target it drains the pending queue. Saves deduped notes to tutored.md.| Command | What it does |
|---|---|
/up:init | Detect & confirm this project's test/lint/typecheck/smoke commands and Definition of Done → .scrum/config.json. |
/up:lean [lite|full|ultra|off] | Dial the always-on lean ladder's intensity (default full); persists as lean_mode. |
/up:sprint plan | close | Plan a sprint via a guided question-bundle; close records velocity and seeds the retro. |
/up:story start <id|desc> | add | Plan + lock a story (brief, estimate, contract); or quick-add one to the backlog. |
/up:done | Navigator review + done-gate, then close the active story. |
/up:refactor <target> | Codegraph-impact-checked refactor with tests green throughout. |
/up:retro | Short retrospective appended to .scrum/retro.md. |
/up:tutor <target> | Tutor-driven deep-understanding session over a story, a file/area, the whole project, or a topic; saves what you learn to tutored.md. |
npx claudepluginhub mthanhlm/ultrapower --plugin upLean workflow plugin for agile-driven development - epic, stories, tasks without the overhead
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation