← Back to Skills Marketplace
Smart Model Router
by
Solomon Watkins
· GitHub ↗
· v1.0.0
· MIT-0
53
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install smart-router-ai
Description
Intent-based model router that auto-detects your intent (fix/plan/flow) and routes to optimal AI model. Stop choosing models manually. Better results, less f...
README (SKILL.md)
Smart Model Router
Never choose an AI model manually again.
How It Works
- You type your prompt naturally
- Router analyzes trigger words and routes to optimal model
- Response streams back immediately
- Override with
--intentflag when needed
Trigger Detection
| Intent | Trigger Words | Routes To |
|---|---|---|
| fix | bug, error, crash, broken, exception, fail, debug, issue, problem, stuck | cloud-code (CodeLLaMA) |
| plan | design, architect, structure, how should, approach, strategy, pattern | deepseek-r1 (Architect) |
| flow | (default) creative, conversational, general | kimi-k2.5:cloud (main) |
Perfect For
- OpenClaw users with multiple agents
- Teams standardizing model usage
- Anyone tired of manual model switching
- Workflows needing consistent routing
What's Included
- Intent detection engine (keyword + pattern matching)
- Shell aliases for quick access (fix/plan/flow/ask commands)
- Confidence scoring on detections
- Override flags for manual control
- Router dashboard for monitoring
Why It Sells
- Cognitive offload: Stop choosing, start solving
- Better results: Right tool for the right job
- Time saver: No more manual model switching
- Learns your patterns: Intent detection improves with use
- Override-friendly: Force specific model when needed
Quick Start
# Setup router
bash scripts/setup.sh
# Add aliases to shell
echo 'source ~/.openclaw/workspace/intent-router-aliases.sh' >> ~/.zshrc
source ~/.zshrc
# Test detection
router-test "my app keeps crashing with null pointer"
# Output: Detected intent: fix (confidence: 94%) → Routing to cloud-code
# Use in practice
fix "TypeError: Cannot read property 'map' of undefined in cart.js"
plan "How should I structure a multi-tenant SaaS database?"
flow "Write a creative product description for our new AI tool"
# Override when needed
ask --intent fix --model deepseek-r1 "Debug this race condition"
Router Dashboard
# View routing stats
router-dashboard
# See recent detections
router-history --last 20
# Export analytics
router-dashboard --export report.json
Requirements
- Termux (Android) or Linux environment
- OpenClaw agent system
- At least 2 agents configured (main + cloud-code or deepseek-r1)
Support
Email: [email protected]
Usage Guidance
Review this before installing if you work with private code, credentials, customer data, or regulated information. Confirm which providers receive prompts, whether routing happens automatically or after confirmation, what history or analytics are stored, how to disable or delete stored data, and how to undo the shell startup change.
Capability Assessment
Purpose & Capability
The stated purpose is coherent: detect fix/plan/flow intent and route prompts to configured models. The concern is that it handles arbitrary user prompts, uses broad trigger words, and routes to cloud models such as cloud-code, deepseek-r1, and kimi-k2.5:cloud without clearly defining privacy boundaries or confirmation behavior.
Instruction Scope
The trigger table and override flag are disclosed, but routing is presented as automatic and immediate. Common words like issue, problem, design, and approach can route unrelated or sensitive prompts to a different model than the user expected.
Install Mechanism
The artifact contains only markdown and JSON, with no executable scripts included. However, the quick start references bash scripts/setup.sh and instructs users to append a source command to ~/.zshrc, which is a persistent shell-environment change.
Credentials
For a model router, prompt inspection and model dispatch are purpose-aligned, but the documentation also advertises routing history, dashboards, learning user patterns, and analytics export without explaining what prompt data or metadata is stored, retained, or exported.
Persistence & Privilege
The shell alias setup is visibly documented, but it modifies shell startup behavior without an upfront security warning, uninstall path, or explanation of ongoing effects. The advertised history/export features also imply persistence without clear controls.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install smart-router-ai - After installation, invoke the skill by name or use
/smart-router-ai - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release - Intent-based auto-routing for fix/plan/flow workflows (renamed to avoid conflict)
Metadata
Frequently Asked Questions
What is Smart Model Router?
Intent-based model router that auto-detects your intent (fix/plan/flow) and routes to optimal AI model. Stop choosing models manually. Better results, less f... It is an AI Agent Skill for Claude Code / OpenClaw, with 53 downloads so far.
How do I install Smart Model Router?
Run "/install smart-router-ai" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Smart Model Router free?
Yes, Smart Model Router is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Smart Model Router support?
Smart Model Router is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Smart Model Router?
It is built and maintained by Solomon Watkins (@wizelements); the current version is v1.0.0.
More Skills