rune
Less skills. Deeper connections. 67-skill engineering mesh for AI coding agents โ TDD-first workflows, security gates and project memory. Native in Claude Code, Codex, Cursor, Grok Build and Gemini CLI; works in any Agent Skills runtime. MIT.
Install / Use
npx skills add Rune-kit/runeInstalls into whichever agent you are using.
Other
Other agent config
Quality Score
Category
SecuritySupported Platforms
Our assessment of rune
rune scores 88/100 on our quality scale, 566th of 1,098 Security skills we index.
Its Other is 19 KB long, well organised into 12 sections with 15 code examples: a thorough specification that gives an agent plenty to work with.
It has 94 GitHub stars, a meaningful sign that others use it.
Maintenance, license and trust
- The repository was last updated today, so rune is actively maintained.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit โ read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.
Automated pattern scan on 2026-11-22. It catches known dangerous patterns, not every risk โ read a skill before letting an agent act on it.
rune compared with similar skills
All 4 of these similar skills score higher than rune; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| rune (this skill)by Rune-kit | 88 | 94 | today | Other |
| competitor-profilingby coreyhaines31 | 100 | 60.1k | 11d ago | SKILL.md |
| reverse-engineer-marketingby coreyhaines31 | 100 | 60.1k | 11d ago | SKILL.md |
| ui-ux-pro-maxby nextlevelbuilder | 100 | 149.0k | 12d ago | SKILL.md |
| design-isby thedotmack | 100 | 116.8k | 9d ago | SKILL.md |
Frequently asked questions
- How do I install rune?
- Run
npx skills add Rune-kit/rune. The install tabs above show the steps for each supported agent. - Which AI agents does rune work with?
- It is written for Claude Code, Gemini CLI, Cursor and OpenAI Codex, as a Other file. Other agents that read the same format can often use it too.
- Is rune safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is rune still maintained?
- The repository was last updated today, so rune is actively maintained.
Skill content
View source on GitHubnpx skills add rune-kit/rune
<sub>One command, ~80 agents: installs every Rune skill into the agents you pick, via <a href="https://skills.sh/rune-kit/rune">skills.sh</a>. For hooks, subagents and the full mesh, use the native install for your agent below.</sub>
<!-- โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ DEMO GIF PLACEHOLDER โ replace before launch โ โ Target: assets/demo.gif, โค 60 s, โค 8 MB, 1280ร720 โ โ Story: same task (e.g. "refactor this 450-line component") โ โ run side-by-side without Rune vs with Rune; end on the diff + โ โ token counter. Then uncomment the block below. โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ <p align="center"> <img src="assets/demo.gif" alt="Rune demo โ the same task with and without Rune" width="100%"> </p> -->Why Rune?
Most skill collections are either hundreds of isolated prompts that never talk to each other, or rigid pipelines where one failed step stops everything.
Rune is a mesh โ 67 skills with 248 connections + 46 signals across a 5-layer architecture. Every declared outbound call is acknowledged by its target, so workflows route around failure instead of stalling.
Pipeline: A โ B โ C โ D (B fails = stuck)
Mesh: A โ B โ C (B fails = A reaches C via DโE)
โ โ
D โ E โ F
Rune doesn't make your model smarter โ it already knows how to code. Rune makes it disciplined: TDD by default, quality gates before commit, evidence before "done", and memory across sessions. The more complex the task, the bigger the gap.
| Measured on Claude Code (10 standardized tasks) | Tokens | Cost | Time | |---|---|---|---| | Refactor a 450-line component | โ62% | โ17% | โ32% | | Full feature (auth + API + tests) | โ36% | โ29% | โ31% | | Average across all 10 tasks | โ16% | โ6% | โ9% |
<details> <summary>Full 10-task breakdown and method</summary>| # | Task | Diff | Tokens | Cost | Time | Correct | |---|------|------|--------|------|------|---------| | 1 | Zod Validation | Easy | -9% | -28% | -32% | โ โ โ | | 2 | Fix N+1 Query | Easy | +12% | +25% | +3% | โ โ โ | | 3 | Cursor Pagination | Med | +12% | +19% | -9% | โ โ โ | | 4 | Security Review | Med | +13% | +32% | +3% | โ โ โ | | 5 | Rate Limiting | Med | +12% | +5% | +5% | โ โ โ | | 6 | Refactor Component | Med | -62% | -17% | -32% | โ โ โ | | 7 | Dark Mode (6 files) | Hard | ~0% | +10% | -7% | โ โ โ | | 8 | DB Migration | Hard | +52% | +11% | +49% | โ โ โ | | 9 | Memory Leak Debug | Hard | +13% | +28% | -2% | โ โ โ | | 10 | Full Auth System | Complex | -36% | -29% | -31% | โ โ โ |
Method: Claude Code CLI headless mode (claude -p --output-format json), each task run once without and once with Rune against the same fixture code; correctness checked by pattern-based evaluation. Simple tasks can cost slightly more (the gates do real work); the savings show up as tasks get bigger.
Install
Pick your agent. Every route gives you the same 67 skills.
Claude Code โ native plugin (skills, subagents, hooks):
/plugin marketplace add rune-kit/rune
/plugin install rune@rune-kit
On Claude Code 2.1.275+ the one-step form also works: /plugin install rune --marketplace rune-kit/rune. Then run npx @rune-kit/rune setup once to wire the lifecycle hooks.
OpenAI Codex โ native plugin (skills, .codex/agents roles, synchronous hooks):
npx @rune-kit/rune setup --here --platform codex
Review and trust the new hooks with /hooks in Codex.
Grok Build โ Grok reads Rune's plugin, skills and hooks natively:
grok plugin install rune-kit/rune --trust
# or browse it first: grok plugin marketplace add rune-kit/rune && grok plugin install rune --trust
Cursor โ Cursor plugin (skills, agents, /rune command) is in the repo under .cursor-plugin/; until the Marketplace listing is live, compile into your project:
npx @rune-kit/rune init --platform cursor
Gemini CLI โ extension (skills):
gemini extensions install https://github.com/rune-kit/rune
Everything else โ compile Rune into your agent's native skills folder:
npx @rune-kit/rune init --platform <name> # grok ยท cline ยท copilot ยท windsurf ยท agents ยท โฆ
Platform support
| Platform | How to install | Skills | Lifecycle hooks | Subagents |
|---|---|:-:|:-:|:-:|
| Claude Code | /plugin install rune@rune-kit | 67 | โ
native | โ
|
| OpenAI Codex | .codex-plugin + rune setup --platform codex | 67 | โ
native | โ
spawn_agent |
| Grok Build (xAI CLI) | grok plugin install rune-kit/rune --trust ยท or --platform grok โ .grok/skills/ | 67 | โ
native .grok/hooks/ via rune hooks install --platform grok (beta) | โ
(beta) |
| Cursor | Cursor plugin (.cursor-plugin/) ยท or --platform cursor โ .cursor/skills/ | 67 | โ
native .cursor/hooks.json via rune hooks install --platform cursor (beta) | โ |
| Grok Bot | Cursor plugin (Cursor Marketplace listing pending) | 67 | โ | โ |
| Gemini CLI | gemini extensions install โฆ/rune ยท or --platform gemini | 67 | โ
native .gemini/settings.json via rune hooks install --platform gemini (beta) ยน | โ |
| GitHub Copilot (CLI / VS Code agent mode) | --platform copilot โ .github/skills/ | 67 | โ | โ |
| Cline | --platform cline โ .cline/skills/ | 67 | โ | โ |
| Kiro ยท Amp ยท Goose ยท Factory ยท Junie ยท Trae ยท Roo | --platform agents โ .agents/skills/ ยท or npx skills add rune-kit/rune | 67 | โ | โ |
| Windsurf ยท Antigravity ยท OpenCode ยท Qoder ยท Qwen Code | --platform <name> | 67 | rules (Windsurf, Antigravity) | โ |
| Aider ยท OpenClaw ยท any other | --platform aider / openclaw / generic | 67 | inline constraints | โ |
| Any Agent Plugins 1.0 client | root plugin.json (skills) | 67 | โ | โ |
Every route ships all 67 core skills and the 14 free extension packs. Where a runtime has no native hooks or subagents, Rune compiles the same gates into inline MUST/NEVER constraints and sequential workflows โ same knowledge, runtime-dependent enforcement.
<sub>ยน The Gemini extension's bundled hooks/hooks.json is Claude-format; Gemini skips the events it doesn't know and may log a notice at startup. Skills are unaffected, and rune hooks install --platform gemini writes Gemini-native hooks separately.</sub>
Quick start
/rune onboard # map the project โ CLAUDE.md / AGENTS.md + .rune/ memory
/rune cook "add user authentication with JWT" # full TDD cycle: plan โ test โ build โ gates โ commit
/rune debug "login returns 401 for valid users" # root cause first, then fix
/rune rescue # refactor legacy code behind characterization tests
/rune sentinel # security scan before you commit
/rune audit # 8-dimension project health report
Outside Claude Code, pick a skill from your agent's skill or slash-command menu (compiled skills are named rune-<skill>, e.g. $rune-cook in Codex), or just describe the task โ each skill's description tells the agent when to use it.
What's New (v2.33.0 โ Everywhere)
- Native installs beyond Claude Code. Cursor plugin, Grok Build plugin + marketplace, Gemini CLI extension and an Agent Plugins 1.0 manifest โ plus
npx skills add rune-kit/runefor any of ~80 agents. - 16 compile targets (was 13): new
grok(.grok/skills),cline(.cline/skills) and universalagents(.agents/skillsโ Kiro, Amp, Goose, Factory, Junie, Trae โฆ). - Claude marketplace fixed. Pack ids were invalid (
@rune/ui); they are nowrune-uietc. andclaude plugin validatepasses. rune doctorworks on every directory-per-skill platform (it crashed with EISDIR on Codex, Cursor, Gemini โฆ).- Numbers you can trust. CI now fails if any README, doc or manifest quotes a stale skill / signal / connection / platform count.
- Skills other agents can pick. Every skill description now says when to use it, model examples are current, and Claude-only tool names compile to neutral wording elsewhere.
- Nothing lost in compilation. Every platform now ships each skill's
references/; the 13 largest skills were split into lean SKILL.md + references, so every skill now fits 500 lines (nothing dropped; a test checks every line). - One
AGENTS.mdfor every agent./rune onboard --runtimessets up per-runtime config and hooks; native hooks for Grok Build, Cursor (.cursor/hooks.json) and Gemini CLI (.gemini/settings.json) viarune hooks install --platform grok|cursor|gemini; Codex subagent roles on GPT-6.x. AGENTS.md+llms.txt, a shorter README, and an honest ROADMAP.
Full notes: CHANGELOG.md. Older release notes: CHANGELOG.md.
Free, Pro and Business
| | Free (this repo) | Pro โ $49 lifetime | Business โ $149 lifetime | |---|---|---|---| | 67-skill mesh, 14 domain packs, 16 platforms | โ | โ | โ | | Hooks: preflight, sentinel, completion-gate, quarantine | โ | โ | โ | | Context watch | tool-call counter | real-% context cockpit, burn-rate ETA, auto state save before compaction | โ | | Autopilot โ approve a plan once, it executes and resumes across sessions | โ | โ | โ | | Domain packs | 14 engineering | +9: product, sales, data science, support, growth, media, personal brand, e-commerce, Vietnam | +4: finance, legal, HR, enterprise search + cross-department orchestrators |
One-time payment, no subscription. Card/PayPal (Polar) or Vietnamese bank transfer (SePay).
<p align="center"> <a href="https://rune-kit.github.io/rune#pricing"><img src="assets/pro-upgrade.svg" alt="Rune Pro โ Autopilot + Context Intelligence ยท $49 lifetime" width="100%"></a> </p> <p align="center"> <a href="https://rune-kit.github.io/rune#pricing"><img src="https://img.shields.io/badge/Get_Pro-%2449_lifetime-a855f7?style=for-the-badge" alt="Get Rune Pro"></a> <a href="https://rune-kit.github.io/rune#pricing"><img src="https://img.shields.io/badge/Get_Business-%24149_lifetime-orange?style=for-the-badge" alt="Get Rune Business"></a> </p> <details> <summary>What's inside Pro and Business</summary>Context Intelligence (Pro). On Claude Code, Pulse runs as a native statusline; on Codex it runs as the Pro watcher/HUD next to Codex's own footer.
Truncated for display โ read the full file on GitHub.
Related Skills
competitor-profiling
60.1kWhen the user wants to research competitors into dated profiles, or asks what a rival charges, claims, or is doing, or whether a conclusion about them is fair: positioning, pricing, features, reviews, SEO, observed facts kept apart from inference. For comparison pages, see competitors.
reverse-engineer-marketing
60.1kWhen the user wants to know how another company grew and what to copy, or asks what a competitor did and when: a dated log of their moves against their traffic from public data, archives, job posts, reviews, and their funnel, compared with the user's data, ending in plays routed to other skills.
ui-ux-pro-max
149.0kUI/UX design intelligence for web, mobile, and desktop. This skill should be used when designing, building, reviewing, or fixing interfaces, including pages, components, design systems, accessibility, interaction, responsive layout, typography, color, charts, and stack-specific UI implementation.
design-is
116.8kAudit a design against Dieter Rams' ten "Good design is..." principles, then hand off a /make-plan prompt for one of three outcomes โ new design, refine design, or redesign
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit โ see the Safety scan above for what the skill file itself contains.
