v2.15.0 — 63 skills • second opinion • MIT
A mesh ecosystem for AI coding assistants. Skills call each other bidirectionally, forming resilient workflows that adapt when things go wrong.
Most skill ecosystems are either too many isolated skills or rigid pipelines. Rune is neither.
A → B → C → D
✖ B fails = stuck
Linear. If one step fails, everything stops.
A → HUB → C
✖ HUB fails = stuck
Central bottleneck. Single point of failure.
A ↔ B ↔ C
↕ ↕
D ↔ E ↔ F
✔ B fails = A→D→E→C
Resilient. Skills route around failures automatically.
10 coding tasks. Claude Code with Rune vs without. Same model, same prompts, real metrics.
| Task | Difficulty | Tokens | Cost | Duration | Tools |
|---|---|---|---|---|---|
| Refactor 450-line component | Medium | -62% | -17% | -32% | -27% |
| Full feature (auth + API + tests) | Complex | -36% | -29% | -31% | -27% |
| Add Zod validation | Easy | -9% | -28% | -32% | 0% |
| Dark mode across 6 components | Hard | ~0% | +10% | -7% | -6% |
Rune doesn't make Claude smarter — Claude already knows how to code.
Rune makes Claude disciplined. The more complex the task, the more discipline matters.
Each layer has a clear responsibility. Skills only call downward (with documented exceptions).
Meta-enforcement. Routes every action to the correct skill.
cook • team • launch • rescue • scaffold
plan • debug • fix • test • review • sentinel • design • ba • docs • and 18 more
research • problem-solver • verification • git • session-bridge • and 20 more
Domain-specific: trading, SaaS, mobile, AI/ML, gamedev, Zalo, and 8 more
Orchestrators coordinate the mesh into end-to-end workflows.
/rune cookBuild any feature with TDD discipline
/rune rescueRefactor legacy code safely
/rune teamParallel multi-agent execution
Master plan + phase files. One phase per session. Any model executes accurately. Resumes across sessions automatically.
12 biases checked before every analysis. Reversibility filter, pre-mortem, MECE decomposition. Structured reasoning, not gut feelings.
Every skill auto-selects the optimal model. Haiku for scanning, Sonnet for code, Opus for architecture. ~$0.05-0.15 per feature.
215+ connections. If debug fails, problem-solver takes over. If scout fails, research + docs-seeker route around. Max depth 8, max 2 visits.
sentinel (OWASP, secrets, deps) + sast (static analysis) + integrity-check (prompt injection detection). Gates, not suggestions.
Same 63 skills compile to Claude Code, Cursor, Windsurf, Antigravity, OpenAI Codex, OpenClaw, and generic. Zero knowledge loss across IDEs.
Native Plugin
# Add marketplace (one-time)
claude plugin marketplace add rune-kit/rune
# Install plugin
claude plugin install rune
Full mesh: subagents, hooks, adaptive routing, analytics.
Compiled Rules
npx @rune-kit/rune init
Auto-detects platform. Compiles all 63 skills to native rule format.
ClawHub Registry
clawhub install rune-kit
63 skills as unified mesh package. Works with any OpenClaw agent.
From first install to advanced mesh orchestration — everything in one place.
Skills don't run in isolation. They call each other, forming resilient workflows that adapt when things go wrong.
cook → plan → test ✗ → debug → fix → test ✓ → review → sentinel → git
Autopilot wraps cook in autonomous mode with stricter quality gates. Domain packs register as Called By ← cook — same orchestrator, no config needed.
Business includes Pro because it depends on Pro data. Finance can’t forecast without sales pipeline. Legal can’t audit without product specs. That’s why they’re bundled.
The full development workflow — plan, build, test, review, deploy — runs entirely on Free. You never hit a paywall while coding.
Pro skills plug into the existing mesh via cook and team. Same orchestrator, same quality gates, new department capabilities.
Finance pulls from sales pipeline. Legal checks product specs. Enterprise search indexes support KB. That’s 13 cross-tier connections — and why Business includes Pro.
See what your AI assistant looks like when it understands every department — not just engineering.
launch-product, quarterly-review, compliance-audit, customer-lifecycle — L1 orchestrators that coordinate entire business operations across departments.
40 domain signals wire departments together. Sales data flows to finance. Product specs feed legal review. No manual context passing.
3 governance templates (startup, mid-size, enterprise) with teams, roles, approval flows, and sentinel enforcement levels.
Real cross-department workflows that run with a single prompt.
> Run quarterly planning for Q3
Autopilot executes your plan phase by phase — self-reviewing, regression-checking, and committing. Stricter quality gates than manual mode, because no one’s watching.
When no human is watching, autopilot compensates with stricter verification than manual mode.
cook gets the job done. autopilot gets it done while you sleep.
Included in Pro — $49 lifetime
The full development mesh is free forever. Pro and Business add department-specific skills for teams that ship products.
Solo dev or open-source contributor? Free has everything you need. 63 core skills, 14 domain packs, all platforms.
Shipping a product with a team? Pro adds Product, Sales, Data Science, Support, and Growth Marketing workflows to your AI assistant.
Running a company? Business adds Finance, Legal, HR, and Enterprise Search on top of Pro.
Everything you need to build, test, deploy, and maintain software.
Best for: individual developers, open-source projects, learning AI-assisted coding.
Business department skills that turn your AI assistant into a cross-functional teammate.
Best for: product teams, startups, SaaS companies shipping features to real users.
💰 14-Day Money-Back Guarantee — no questions asked
Full department coverage — every business function, one purchase.
What’s Included:
💰 14-Day Money-Back Guarantee — no questions asked
| Capability | Free | Pro $49 | Business $169 |
|---|---|---|---|
| Core dev skills (63) | ✓ | ✓ | ✓ |
| Domain packs (14) | ✓ | ✓ | ✓ |
| Multi-platform compiler | ✓ | ✓ | ✓ |
| Security gates & SAST | ✓ | ✓ | ✓ |
| Autopilot — autonomous multi-session | — | ✓ | ✓ |
| Product Management | — | ✓ | ✓ |
| Sales Enablement | — | ✓ | ✓ |
| Data Science | — | ✓ | ✓ |
| Customer Support | — | ✓ | ✓ |
| Business memory layer | — | ✓ | ✓ |
| MCP connector templates | — | ✓ | ✓ |
| Finance Ops | — | — | ✓ |
| Legal & Compliance | — | — | ✓ |
| HR & Recruiting | — | — | ✓ |
| Enterprise Search | — | — | ✓ |
| Skill Forge (build custom skills) | ✓ | ✓ | ✓ |
| Business-Exclusive | |||
| L1 Business Orchestrators (4) | — | — | ✓ |
| Cross-domain signals (40) | — | — | ✓ |
| Org governance templates (3) | — | — | ✓ |
| Cross-department automation | — | — | ✓ |
| Priority support | — | — | ✓ |
14 domain-specific packs included with Rune. Install what you need.
Choose your payment method
1,190,000 VND (~$49 USD)
Transfer the exact amount with the code below
Open your banking app → Scan QR or transfer manually → We'll detect it automatically
A GitHub repo invitation is on its way to
noreply@github.com — usually within 1 minute
Didn't get the invite? Check spam, or email claude@theio.vn
$49 USD
You'll be redirected to Polar's secure checkout page
Lời mời truy cập repo đã được gửi đến
noreply@github.com — thường trong vòng 1 phút
Chưa nhận được? Kiểm tra spam, hoặc email claude@theio.vn