Development plugins

Yeet builds and deploys your project, making it available online.

1 command · 2 agents

A disciplined, self-verifying feature-dev loop: every commit unit gets dual-model review and a reconcile pass, tests are the ground truth, and you decide what lands (no auto-commit). Stack-agnostic via config plus auto-detect.

1 skill · 1 command · 1 agent

Appcircle skills covering mobile CI/CD, build, code signing, distribution, publishing, and testing, with MCP connectivity to the Appcircle platform.

2 skills · 1 MCP server

Skills for working with a project's intent — what it is for, as distinct from what it does or how it is built. Recover it into an INTENT.md, assess a project against it, and test which design choices it actually forces.

3 skills

Create and validate native Claude Code artifacts with conservative generators and deterministic checkers.

3 skills

Iterativer Loop mit 4 Rollen und 5 Standard-Adaptern (dummy, local_shell, notebooklm_cli, codex_rescue, openai_chat). State (PLAN/TODO/LOG/DECISIONS/RESULTS) liegt projekt-lokal unter .loop/<run-id>/.

2 skills · 1 command

Bootstrap — interview for persona, then generate a minimal personalized workbench (/workbench-kit:generate-workbench).

3 skills

Development lifecycle — autonomous task execution (Linear), PR CI fixes, and PR review comment resolution

3 skills

Complete Neam language skill for Claude Code — covers 13 data types, agents, claw/forge agents, multi-agent orchestration, RAG, skills, tools, MCP, guards, budgets, OOP, modules, deployment, and 100+ built-in functions

3 skills

Inject architectural thought into one-shot LLM coding: the archkit MCP server (31 tools, CGR relay prompts, @archkit: resources), four guardrail hooks (SessionStart, Stop, PostToolUse, UserPromptSubmit), and the /archkit-init wizard.

1 skill · 1 MCP server · hooks

Run codeflash as a background agent to optimize code for performance (Python, Java, JavaScript, TypeScript)

2 skills · hooks

Drop documentation-drift + GitHub-release→changelog automation into any repo: reusable GitHub Actions, AGENTS.md conventions, and (for Next.js + Prisma apps) a Help Centre, Dev Docs viewer, and changelog ingest route.

1 skill · 2 commands

Donne a Claude les yeux d'un joueur Minecraft : captures, isolation d'un sujet sur fond uni, cadrage automatique d'une entite, lecture des inventaires et des lores, detection des regressions visuelles. Necessite le mod Fabric mcbridge dans un client Minecraft 26.1.2.

1 skill · 1 command · 1 MCP server

Plan, create and claim zero-cost AI agent teams on Central City over MCP (no account needed), plus OAuth access to your workspace.

1 skill · 2 MCP servers

code review toolkit, uses multi sub-agent based workflow to review a local branch or a gitlab merge request

1 command · 2 agents

backend development toolkit, containing supporting skills for tdd, writing use cases and defining api specs

2 skills · 1 agent

Padrao organizacional e de performance para qualquer projeto backend, independente de framework. Estrutura, camadas, nomenclatura, performance, resiliencia, observabilidade e portoes executaveis.

2 skills · 1 agent

Create and maintain an audience-aware HTML research notebook with daily work, ranked assumptions, next steps, and ASTRA-linked results when available.

3 skills

A plugin for spec-driven development

3 commands

Temporal Python workflow skills for retry policies, testing, replay safety, and deterministic deployments.

3 skills

NeevCloud agent skills bundle: the SDK skill for creating sandboxes, running code and processes, and exposing a port to get a public preview URL, the CLI skill for signing in, choosing an organization and project, and managing sandboxes from the command line, and the sandbox MCP skill for connecting an agent to a sandbox as native tools.

3 skills

Personal feature-development workflow chain extending superpowers with research + multi-subagent review phases.

3 skills

[alpha] /skill-eval — Claude Code skill の効果を Copilot SDK で A/B 検証する CI を生成する。

3 commands

Measure a website's brand identity (color, type, spacing, motion) from the live page, then learn its feel.

2 skills · 1 command

Stop the agent from grepping its way through a large repo. Augments codebase-memory-mcp (>= 0.11.0) with what it does not ship: a discovery ladder with silent-truncation detection, LSP as the tiebreaker, narrow-window editing, verification by test count, a sweeping-change playbook for migrations that outgrow one context, and hooks that enforce all of it -- including auto-approval of read-only graph queries so they do not prompt in plan mode. Hooks run on macOS, Linux and Windows without Git Bash.

2 skills · hooks

Optional MCP servers for gamedev-forge: Blender (drive a running Blender), Godot (run and inspect projects), and Context7 (current engine and library docs). No API keys are needed.

3 MCP servers

Claude Code 插件创建向导 - 引导生成 agents、commands、skills、hooks

1 skill · 1 command · 1 agent

delegate-task hands one task to another agent CLI and reports what it changed and the tokens it used. progress-report keeps a live progress report for a Markdown plan or a Spec Kit feature. planning-majlis plans software or a feature with a committee of expert roles run as subagents.

3 skills

開発マシン／新環境を冪等・非破壊にオンボーディングする。エージェント本体（Claude Code / Codex CLI）の導入、版管理された dotfiles の反映（配布方式の選び分け付き）、シークレット混入を commit 前に止めるグローバルガードの 3 本立て。

3 skills

プロジェクトの品質規約を正本として持ち、リポジトリを監査して是正をタスク化する。docs サイトと PyPI リリースの公開 CI も、CI を割らない安全な guard 付きで整える（Python + uv）。

3 skills