Skills, subagents, slash commands, hooks, settings, plugin bundles and MCP servers that install into Claude Code. 388,310 indexed, measured the same four ways as everything else here.
AI-Driven Development framework: Ideate -> Research -> human decision gate in the browser -> Execution. Research produces a themed HTML gate where every option can be Applied, Inferred, or Innovated on; submitting compiles the choices into a binding markdown contract that drives the build.
★not rated 2 2mo agoA
tokens not measured
originalMIT
Turns Claude Code wait time into workout time. When a long command runs (build, install, tests, docker...), you get told to do push-ups or squats instead of doomscrolling.
★not rated 2 2mo agoA
tokens not measured
originalMIT
A project-status scanner and workflow recommender, described in Chinese, for guiding development work through a ten-step process. It gathers information about Git, progress, specifications, audits, and changes, then proposes a workflow for user approval.
A guide to communication in Chinese workplaces, including mixed Chinese-English translation, internet slang, socially tactful wording, and implied meanings.
Agentic SDLC for Claude Code — Claude facilitates writing specs and designing architecture, then autonomously builds implementation plans, implements and tests the product. Driven by structured /agn: skills with built-in review gates and document consistency.
★not rated 2 4mo agoA
tokens not measured
originalApache-2.0
Open-source mirror of the qship pipeline — Jira ticket to production PR in one slash command. Hook-enforced, resumable, parallel/epic-mode capable, with optional Codex CLI integration for diversified model review.
★not rated 2 4mo agoA
tokens not measured
originalMIT
Tracks every Claude Code session as a GitHub Issue in the repository you are currently working in. Creates a new issue per session, appends every prompt and response as comments.
★not rated 2 6mo agoA
tokens not measured
originalMIT
Cross-vendor LLM advisors for Claude Code: ask codex or a local ollama/lmstudio model, or run a blind independent review across several, from inside a session. Read-only by default on adapters whose boundary is verified; debate/council are experimental.
★not rated 2
changed 23d agoA
tokens not measured
originalMIT
Play a game in your terminal while Claude thinks. Auto-pauses when Claude replies, resumes on your next prompt. Five titles: Break-In, Skyrun, Jetwash, Astros and Jaywalk.
★not rated 2 3mo agoA
tokens not measured
originalMIT
Real-time HUD statusline for Claude Code with API balance monitoring. Shows context health, tool activity, agent tracking, todo progress, git status, and DeepSeek/OpenAI/Anthropic balance tracking with low-balance warnings.
★not rated 2 2mo agoA
tokens not measured
originalMIT
A Korean-language safety loop for non-developers who ask an AI to keep repairing a task until it works. Its description says it currently supports a single repair mode with 14 safety guardrails enabled by default.
Describe what you want. Read one page. Approve four times. Get a product that does what the page says, proven live. One command: /compass. Four stages: Idea, Plan, Build, Go live. Fresh reviewers test every promise inside every stage.
★not rated 2
changed 2d agoA
tokens not measured
originalMIT
Use this platform skill whenever a bound Skill contains scripts or an agent must create files in its isolated application workspace.
★not rated 2 27d agoA30 tokens
originalApache-2.0
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: