By marcrabadan
An on-demand team of SDLC experts for Claude Code — the core six (business analyst, product owner, software architect, developer, QA engineer, DevOps engineer) plus specialists for security, frontend, data, ML/AI, and UX — and a /new-feature orchestrator that runs an idea through the lifecycle, pulling in the specialists a feature warrants. Use to consult one expert or run a whole feature end to end.
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 shared library of Claude Code skills. A repo where a team captures how it does things — processes, conventions, roles, taste — once, in a format Claude Code reads and applies consistently across every project.
It has three parts:
skill-creator — the meta-skill that is the pattern for creating new skills. Use it to scaffold, review, classify, and validate any new skill.memory — a versioned memory ledger that records the plans, decisions (from any role, not just the architect), implementations, and artifacts the experts produce, each with a status from a closed set — pending → accepted | rejected | rolled-back (plus superseded) — so the record survives across sessions and changes can be rolled back. pending is a proposal awaiting your call, not approval to act.On top of these, an experimental harness mode is taking shape — a source-of-truth authority model, per-project memory, explicit stop conditions, and a deterministic harness validator — so praxis can give agents a reliable operating environment, not just skills. It's additive: no existing skill or command changes.
Just want to use it? Jump to Install & integrate for Claude Code, Cursor, IntelliJ, and Codex. Want to see the output first? Browse examples/ for sample transcripts. Otherwise: PMs, designers, stakeholders read from the top; developers writing or shipping a skill skip to the Developer guide.
praxis is a factory for Claude Code skills. A "skill" is a small package of instructions that teaches Claude Code how to do one specific thing your team's way — act as a particular SDLC expert, follow a review checklist, write requirements, design an architecture, and so on.
Every time someone teaches Claude a useful pattern, that knowledge is captured once here and becomes available to everyone. It is not a product repo — it holds the instructions Claude uses when working on product repos, not the products themselves.
Pick the path that matches how you work. The experts are Claude Code native; the same personas are also generated for Cursor, IntelliJ, and OpenAI Codex so you can use them from those tools too.
Prerequisites: Claude Code for the primary path. Python 3 is only needed if you want to author skills with the factory tooling (the validators use the standard library — no extra packages to install).
To pull the experts into a different repo, praxis ships as a Claude Code plugin marketplace with two plugins. Run this from inside the target project:
/plugin marketplace add marcrabadan/praxis
/plugin install praxis@praxis # the twelve SDLC experts + /new-feature + /review-changes
/plugin install skill-factory@praxis # optional: author your own skills (skill-creator + /validate-skills)
As plugins, commands are namespaced under the plugin name — /praxis:architect, /praxis:new-feature, /skill-factory:validate-skills, and so on.
Want just one skill, with no plugin? Copy the folder, or use the export target from inside this repo:
make export SKILL=software-architect TO=../my-product-repo # → ../my-product-repo/.claude/skills/software-architect
# or by hand, including for user-wide scope:
cp -R .claude/skills/software-architect ~/.claude/skills/
The twelve experts ship as Cursor project rules (auto-attaching) plus commands. Copy the generated .cursor/ directory into your project root:
cp -R integrations/cursor/.cursor <your-repo>/
cp AGENTS.md <your-repo>/ # optional: Cursor reads AGENTS.md natively
Personas auto-attach when your request matches them; invoke one explicitly with /architect, /developer, /security, /frontend, … or run /new-feature and /review-changes. Details: integrations/cursor/.
Ships as Junie guidelines + persona guides + ready-to-paste prompts. Copy the generated .junie/ directory into your project root:
cp -R integrations/intellij/.junie <your-repo>/
cp AGENTS.md <your-repo>/ # JetBrains AI Assistant reads AGENTS.md natively
npx claudepluginhub marcrabadan/praxis --plugin praxisComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
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
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.