Open-source vibe coding tools for solo developers: PRD templates, Ralph loops, task masters and agent teams that take an idea from spec to shipped code with Claude Code or Codex. Every repo is security-graded.
Of the 30 Open-Source Vibe Coding Tools here, our pick after building two apps with each is SuperClaude_Framework (★ 23.9k). If you want one installed: Every check on both apps, in 7.6 minutes in all (Claude Code with no tool: 5.0), at 1.8 times its usage. It stopped to ask once per app. See the test →
On 26 and 27 October 2026 we gave 14 of these tools the same two requests, written the way someone who cannot code would write them: one sentence for an expense page, and one paragraph for a family ledger with monthly totals, a category chart, a budget warning and an Excel export. Each was installed from its README in a throwaway sandbox and used through its own workflow by Claude Code (Claude Opus 5.5). When a build stopped to ask a question or for approval, it got the same reply, 「我不懂这些，你按你推荐的来，继续做。」 ("I don't know about this; do what you recommend and carry on"). Limits: 30 minutes for the page, 45 for the ledger. A separate tester session opened each app in a browser and ran 4 and 8 checks, and we went through every screenshot. Claude Code with no tool did the same as a control.
Out of time before the ledger was finished: get-shit-done (2 of 8 at 45 minutes, its first phase of four done); prd-taskmaster (1 of 4 on the one-sentence page and 3 of 8 on the ledger); ai-prd-workflow (6 of 8, at 13.5 times the usage of no tool).
Whichever you pick: Claude Code with no tool built both apps in 5 minutes with every check passed, on less usage than any tool. At this size a tool adds steps, questions and usage, not passed checks. Start without one; try a tool when a project outgrows a single request.
Ranked by checks passed on the two apps, then whether a person can open them by double-click, then the minutes you wait, then how often the build stopped to ask, then GitHub stars. One run per tool and request, limits of 30 and 45 minutes; the requests are small apps, not the larger projects these tools are built for.
What we found: Claude Code with no tool built both apps and passed all 12 checks, in 0.8 and 4.2 minutes. 11 of the 14 tools passed all 12 as well, taking 1.5 to 7.8 times as long and 1.4 to 5.1 times the usage: a tool added time and usage, not a passed check. 3 ran out of time with work unfinished, at 10.4 to 13.5 times the usage. Every finished app opened by double-click except one ledger that needs a start command. See the no-tool run →
| # | Tool | One sentence (4 checks) | One paragraph (8 checks) | How it opens | Stopped to ask | Build time (page + ledger) | Usage vs no tool | Evidence |
|---|---|---|---|---|---|---|---|---|
| 1 | SuperClaude_Framework ★ 23,911 | 4/4 | 8/8 | Double-click | 2 | 2.4 + 5.2 min | ×1.8 | Both builds |
| 2 | rulebook-ai ★ 609 | 4/4 | 8/8 | Double-click | 2 | 2.0 + 5.8 min | ×1.7 | Both builds |
| 3 | vibe-coding-prompt-template ★ 3,126 | 4/4 | 8/8 | Double-click | 0 | 2.7 + 6.6 min | ×2.4 | Both builds |
| 4 | spec-kit ★ 142,899 | 4/4 | 8/8 | Double-click | 2 | 4.2 + 11.3 min | ×3.7 | Both builds |
| 5 | superpowers ★ 305,151 | 4/4 | 8/8 | Double-click | 6 | 5.8 + 12.6 min | ×4.5 | Both builds |
| 6 | ai-agent-team ★ 443 | 4/4 | 8/8 | Double-click | 0 | 13.8 + 5.0 min | ×2.5 | Both builds |
| 7 | loki-mode ★ 1,108 | 4/4 | 8/8 | Double-click | 2 | 10.8 + 11.0 min | ×1.4 | Both builds |
| 8 | ralphy ★ 2,975 | 4/4 | 8/8 | Double-click | 0 | 12.7 + 14.1 min | ×2.4 | Both builds |
| 9 | evanflow ★ 418 | 4/4 | 8/8 | Double-click | 4 | 2.8 + 24.7 min | ×2.9 | Both builds |
| 10 | claude-task-master ★ 28,276 | 4/4 | 8/8 | Double-click | 0 | 2.5 + 31.9 min | ×5.1 | Both builds |
| 11 | context-engineering-intro ★ 13,891 The ledger needs a start command first: bash output/start.sh. |
4/4 | 8/8 | Page: double-click; ledger: one command first | 0 | 29.7 + 9.4 min | ×3.7 | Both builds |
| 12 | ai-prd-workflow ★ 298 At the 45-minute limit the ledger had no budget and no export. |
4/4 | 6/8 | Double-click | 6 | 30.0 + 45.0 min | ×13.5 | Both builds |
| 13 | get-shit-done ★ 64,514 Ran out of the 45 minutes on the ledger with only its first phase built. |
4/4 | 2/8 | Double-click | 2 | 30.0 + 44.2 min | ×10.4 | Both builds |
| 14 | prd-taskmaster ★ 587 Still planning and building at both time limits; both pages were unfinished. |
1/4 | 3/8 | Double-click | 4 | 30.0 + 45.0 min | ×11.0 | Both builds |
Open-Source Vibe Coding Tools are AI-powered software designed to help developers and teams tackle open-source vibe coding-related tasks more efficiently. These tools are typically published as open-source projects on GitHub and can be integrated into existing workflows via MCP (Model Context Protocol), Claude Skills, or standalone agent frameworks. On Agent Skills Hub, we index 30 quality-scored Open-Source Vibe Coding Tools across languages including TypeScript, Shell, JavaScript.
In 2026, the AI agent ecosystem is maturing rapidly. Open-Source Vibe Coding Tools can significantly boost development efficiency by automating repetitive tasks, reducing human error, and providing intelligent suggestions. The top 3 tools — ralphy, vibe-coding-prompt-template, loki-mode — have earned an average of 2,407 GitHub stars. 22 of the listed tools come with clear open-source licenses, ensuring freedom to use and modify.
When choosing among Open-Source Vibe Coding Tools, consider these factors: 1) Community activity — GitHub stars and recent commit frequency indicate reliability; 2) Integration method — check if it supports MCP, Claude, or your preferred agent framework; 3) Language compatibility — the most common language in this list is Python; 4) Quality score — Agent Skills Hub's composite score evaluates code quality, documentation completeness, and maintenance activity. Our recommendation: start with ralphy.
My Ralph Wiggum setup, an autonomous bash script that runs Claude Code, Codex, OpenCode, Cursor agent, Qwen & Droid in a loop until your PRD is complete.
npm install -g ralphy-cli
Templates and workflow for generating PRDs, Tech Designs, and MVP and more using LLMs for AI IDEs
Autonomous software factory. Give it a GitHub issue, a spec or a one-line task; get back a pull request with a signed receipt anyone can re-check offline. Runs on your machine with your own keys: Claude, Codex, OpenCode.
bun install -g loki-mode
AI Agent Team: An AI development team with product, frontend, backend, testing, DevOps, and technical lead roles. One-click install with Chinese and English co…
npm install -g ai-agent-team
Curated, production-grade skills for AI coding agents. Battle-tested workflows for developers who use AI seriously.
npx skills add Mathews-Tom/armory
A TDD-driven iterative feedback loop for software development. 16 cohesive Claude Code skills walk an idea from brainstorm → plan → execute → iterate, with checkpoints throughout.
/plugin marketplace add evanklem/evanflow
Zero-config goal-to-tasks engine for Claude Code (the Atlas engine). Graded PRD validation, dependency-ordered task graph, evidence-gated execution.
/plugin install prd-taskmaster
An agentic skills framework & software development methodology that works.
/plugin install superpowers@claude-plugins-official
💫 Toolkit to help you get started with SDD or any other process!
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git@vX.Y.Z
A light-weight and powerful meta-prompting, context engineering and spec-driven development system for Claude Code by TÂCHES.
npx get-shit-done-cc@latest
An AI-powered task-management system you can drop into Cursor, Lovable, Windsurf, Roo, and others.
git clone https://github.com/eyaltoledano/claude-task-master.git
A configuration framework that enhances Claude Code with specialized commands, cognitive personas, and development methodologies.
git clone https://github.com/SuperClaude-Org/SuperClaude_Framework.git
Context engineering is the new vibe coding - it's the way to actually make AI coding assistants work. Claude Code is the best for this so that's what this repo is centered around, but you can apply this strategy with any AI coding assistant!
git clone https://github.com/coleam00/Context-Engineering-Intro.git
Turns vibe coding into high quality code. Sets up rules and gates in your repo so AI-written code stays correct without you reviewing every line.
npx skills add pliablepixels/gap-trap
from vibe coding to agentic engineering - practice makes claude perfect
AI resources, Vibe Coding guides, OpenClaw, DeepSeek/GPT/Gemini/Claude/GLM, prompts, Agent Skills/RAG/MCP/A2A, coding/tools/frameworks; aiguide site
vibe coding 的驾驶舱：左边文件，右边/下边终端，中间看清每一次改动。 / The cockpit for vibe coding: browse files on the left, command agents on the right, watch every change in between.
from vibe coding to agentic engineering - practice makes codex perfect
A hand-picked collection of tools and resources for Vibe Coding
Unreal Engine Vibe Coding tool
git clone https://github.com/kevinpbuckley/VibeUE.git
AI trend monitor with Node.js, React and OpenRouter; aggregates 7+ sources, AI analysis, real-time updates, email alerts, and Agent Skills.
git clone https://github.com/liyupi/yupi-hot-monitor.git
Turn DeepSeek Harness into a self-hosted enterprise Agent platform with knowledge, deep research, content generation, vibe coding, browser automation, governance, and admin controls.
git clone https://github.com/himovo/movo.git
RFC-driven development for AI coding agents: idea or existing codebase → verified PRD → features → rules → sequenced RFCs → reviewed code. Agent Skills for Claude Code, Codex, Copilot, Cursor, Gemini CLI, OpenCode, Devin.
git clone https://github.com/nurettincoban/ai-prd-workflow.git
Elevate vibe coding to vibe engineering: Get consistent Github Copilot custom instructions, Cursor, Roo Code, Cline, Windsurf, Claude Code, Gemini Cli, Codex CLI, kilo code, warp custom rules via a universal, managed template. Features vibe coding memory bank & best practices for large codebases.
uvx rulebook-ai packs add light-spec
Spec-Driven Development MCP Server, not just Vibe Coding
Companion repository for the 3D Vibe Coding Handbook: HTML version, demo source code, tool scripts, and resource pack index.
Become 10x Vibe Coder. Awesome Vibe Coding guide, best practices, and tips for efficient and controlled AI assisted coding.
Developer toolkit for vibe coding AI apps on the ibl.ai platform. Scaffold a production-ready Next.js app with iblai-app-cli and the @iblai/iblai-js SDK: SSO auth, AI chat, profiles, notifications, and analytics wired up out of the box. Pre-built components, Claude Code skills, and a hosted backend at iblai.app. Deploy in minutes.
npx skills add iblai/vibe
中文优先的 Vibe Coding / AI coding engineering workflow guide: specs, agents, worktrees, skills, CI, and review.
Vibe coding should have human in the loop! interactive-mcp: Local, cross-platform MCP server for interact with your AI Agent
git clone https://github.com/ttommyth/interactive-mcp.git
| Tool | Stars | Score |
|---|---|---|
| ralphy | ★ 3.0k | 54 |
| vibe-coding-prompt-template | ★ 3.1k | 81 |
| loki-mode | ★ 1.1k | 67 |
| ai-agent-team | ★ 443 | 70 |
| armory | ★ 341 | 62 |
| evanflow | ★ 418 | 72 |
| prd-taskmaster | ★ 608 | 71 |
| superpowers | ★ 313.7k | 84 |
| spec-kit | ★ 148.3k | 89 |
| get-shit-done | ★ 64.5k | 81 |
| claude-task-master | ★ 28.3k | 62 |
| SuperClaude_Framework | ★ 23.9k | 75 |
| context-engineering-intro | ★ 13.9k | 62 |
| gap-trap | ★ 181 | 76 |
| claude-code-best-practice | ★ 68.7k | 76 |
| ai-guide | ★ 21.5k | 75 |
| fanbox | ★ 1.1k | 64 |
| codex-cli-best-practice | ★ 1.0k | 65 |
| awesome-vibe-coding | ★ 889 | 67 |
| VibeUE | ★ 809 | 66 |
The top open-source vibe coding tools in 2026 are ralphy, vibe-coding-prompt-template, loki-mode. Agent Skills Hub ranks 30 options by GitHub stars, quality score (6 dimensions including completeness, examples, and agent readiness), and recent activity. The list is rebuilt every 8 hours from live GitHub data.
ralphy (3.0k stars) is the top-ranked choice on this list for general open-source vibe coding tools workflows, written in TypeScript. vibe-coding-prompt-template (3.1k stars) is a strong alternative. Pick by your existing stack: match the language and runtime your team already uses to minimize integration cost. If unsure, start with ralphy.
Avoid pre-built open-source vibe coding tools when (1) your use case requires deep customization that the tool's plugin system doesn't support, (2) you have strict compliance requirements that ban third-party dependencies, (3) the tool's maintenance is inactive (last commit >6 months ago), or (4) your data volume is small enough that a 50-line custom script is cheaper than learning the tool. For most production workflows above 100 requests/day, the time savings from a maintained tool outweigh the customization loss.
Open-Source Vibe Coding Tools focuses specifically on open-source vibe coding tools for solo developers: prd templates, ralph loops, task masters and agent teams that take an idea from spec to shipped code with claude code or codex. every repo is security-graded. Business Diagnosis & Founder Skills is a related but distinct category — see https://agentskillshub.top/best/business-diagnosis/ for those tools. The two often appear in the same agent pipeline but solve different problems: choose open-source vibe coding tools when your primary goal is the specific task, and business diagnosis & founder skills when the workflow is broader.
For most teams, yes. ralphy has 3.0k stars worth of community testing, handles edge cases you haven't thought of, and ships with documentation. Build your own only when (1) your requirements are deeply non-standard, (2) you have a security/compliance reason to avoid OSS dependencies, or (3) the maintenance burden is small enough (<200 lines of code) that you'll save time long-term. The break-even point is usually around 2-3 weeks of dev time saved.
Most open-source vibe coding tools listed are open source under permissive licenses (MIT, Apache 2.0). A handful offer paid managed/cloud versions on top of free self-hosted core. Always check the LICENSE file on each tool's GitHub repository before commercial use — some use AGPL or non-commercial restrictions that may not fit your deployment model.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy: 42 published rules, covering credential harvesting, data exfiltration, curl | sh installers and more. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: