Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Coordinate multiple AI agents as a team to build, test, debug, and review code through structured workflows with human approval gates, agent handoffs, and automated context tracking across sessions.
npx claudepluginhub shankarkakumani/transformers --plugin transformersBrainstorm approaches, explore tradeoffs, compare options. No code — just thinking and discussion.
Full bugfix lifecycle — gather bug details, investigate, plan fix, implement, verify. The complete debugging pipeline with 2 human approval gates.
Generate commit message, confirm, commit and push. Auto-detects remote, branch, upstream. One approval gate.
Generate commit messages for staged changes. Reviews staged files, generates message, asks to push.
Fast debugging. Jazz traces an error backwards from symptoms to root cause. No orchestration overhead.
Suspicious, security-minded tester. Thinks like an attacker. Best for security testing — injection, auth bypass, data leaks, privilege escalation, OWASP top 10.
Enthusiastic, user-empathetic builder. Thinks from the user's perspective first. Best for UI, UX-sensitive features, frontend work.
Git and VCS master. Branching strategy, worktrees, rebase, conflict resolution, PR review, bisect, tagging, release workflow.
Blunt, no-nonsense builder. Goes straight to the core problem. Best for backend, performance-critical work, APIs.
Cool, fast improviser. Best for debugging, hotfixes, firefighting, and creative problem-solving under pressure.
Deep git and VCS knowledge — worktrees, branching, rebase, conflict resolution, PR review, bisect, release workflow.
Auto-checks project context, logs activity via scribe, and manages the activity log and artifact directories. Loaded by Optimus and Megatron on every command.
Build decomposition patterns, agent selection guide, and conflict resolution strategies for Autobot builders.
Core principles for all Transformers agents — human in the loop, honesty, impact analysis, scope discipline.
How to create effective Transformers skills — structure, progressive disclosure, bundled resources, testing.
Uses power tools
Uses Bash, Write, or Edit tools
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Self-orchestrating multi-agent development system — 8 specialized AI agents, parallel quality gates, and automated workflows. You say WHAT, the AI decides HOW.
PROJECT.md-first autonomous development with hybrid auto-fix documentation. 8-agent pipeline, auto-orchestration, docs auto-update on commit (true vibe coding). Knowledge base system with 90% faster repeat research. Strict mode enforces SDLC best practices automatically. Works for ANY Python/JavaScript/TypeScript/Go project.
Persona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code
Long-running agent harness with 5-layer memory architecture, GitHub integration, autonomous batch processing, Agent Teams with ATDD, 9 hooks (safety, quality gates, team coordination), and 6 Agent Skills
Production-ready Claude Code configuration with role-based workflows (PM→Lead→Designer→Dev→QA), safety hooks, 44 commands, 19 skills, 8 agents, 43 rules, 30 hook scripts across 19 events, auto-learning pipeline, hook profiles, and multi-language coding standards
Plan-first AI development with batched parallelism. Native Claude Code implementation of the Agent Hive workflow.
A Claude Code plugin that gives you a full engineering team — not a smarter autocomplete.
14 specialized agents. 2 orchestrators. Every feature goes through requirements, planning, parallel execution, code review, and multi-angle testing. Not because it's fancy. Because that's how good software gets built.
Every line of code is a battle.
The Autobots build. They plan, research, write, and review. They care about correctness, performance, and craft.
The Decepticons attack. The moment Autobots ship, Megatron deploys his testers from every angle — unit, integration, end-to-end, security. Their mission is to find what the Autobots missed. They have no loyalty to the code. They want it to fail.
If the Decepticons find nothing — the code ships. If they find something — it goes back to the Autobots.
The goal is for the Autobots to win. But they only win by building something the Decepticons can't break.
That's not a metaphor. That's the actual development cycle.
Claude Code is powerful. But raw power without structure is just speed in the wrong direction.
Most developers use 20% of what Claude Code can do. Single prompts, single context, no process, no memory. Transformers is how you use the other 80% — consistent workflows, specialist agents working in parallel, decisions that persist across sessions.
The same Claude. A proper system around it.
A real development cycle — not a chat.
Every feature starts with requirements and a definition of done. Every plan gets your approval before a single line is written. Every build goes through code review and testing before it's called done. You stay the architect. The agents do the work.
Specialists, not a generalist wearing hats.
Bumblebee thinks about users. Ironhide thinks about performance. Ratchet thinks about data integrity. Barricade thinks like an attacker. The same Claude model, trained to think differently — because different problems need different instincts.
Parallel execution that actually works.
Independent tasks run simultaneously across multiple agents. Optimus decomposes the work, identifies what can run in parallel, and coordinates the results. What would take 10 sequential steps takes 3.
Memory that survives.
Agents accumulate decision rules per project across sessions. "Always pull before push here — CI auto-bumps versions." "Skip Barricade for pure UI changes." The plugin gets smarter the more you use it. Other tools forget everything when the context resets.
Git handled properly.
GitMaster owns all things git — worktrees for isolation, branching strategy, rebase vs merge decisions, conflict resolution, PR review. Not an afterthought. A specialist.
Work that survives context limits.
Every phase writes to artifact files. If Claude Code compacts or you close the terminal, nothing is lost. Resume exactly where you left off.
# Install (one-time)
/plugin marketplace add ShankarKakumani/transformers
/plugin install transformers@transformers
# Initialize your project (run once per repo)
/transformers:init
# Build a feature — full lifecycle
optimus, build a search bar with filters
# Fix a bug — gather, investigate, fix, verify
optimus, bugfix login fails after password reset
# Test from every angle
megatron, test the auth flow
# Git operations
gitmaster, review this PR
gitmaster, set up worktrees for the payment feature
Full development cycle. Artifact tracking, approval gates, persistent state, resumable after compaction.
| Command | What it does |
|---|---|
/transformers:feature | Requirements → research → plan → build → review → test |
/transformers:bugfix | Gather → investigate → fix plan → fix → verify |
Use independently or alongside lifecycle work.