← Back to Skills Marketplace
local-config-model-recommender
by
garygou1024
· GitHub ↗
· v1.0.0
· MIT-0
265
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install local-config-model-recommender
Description
Intelligently recommends optimal AI models based on task requirements. Dynamically reads the user's OpenCLAW configuration and provides context-aware model s...
README (SKILL.md)
Model Recommender
An intelligent model selection assistant that dynamically analyzes your OpenCLAW configuration and recommends the most suitable AI model for your specific task.
Overview
This skill reads your local OpenCLAW configuration to determine available models, then applies keyword-based capability matching to provide intelligent recommendations based on the task at hand.
Supported Models
Global Providers
| Provider | Models |
|---|---|
| OpenAI | gpt-5-pro, gpt-4o, gpt-4o-mini, o3, o1 |
| Anthropic Claude | claude-opus-4.6, claude-sonnet-4.6, claude-haiku-4.5 |
| Google Gemini | gemini-3.1-pro, gemini-3.1-flash, gemini-2.5-flash |
| Mistral | mistral-large, mistral-small |
| xAI | grok-4.20-beta, grok-4-fast |
Domestic (China) Providers
| Provider | Models |
|---|---|
| Alibaba Qwen | qwen3-max, qwen3-vl, qwen3.5-plus, qwen3-coder-plus |
| MiniMax | minimax-m2.5, minimax-m2.1 |
| DeepSeek | deepseek-v3.2, deepseek-chat, deepseek-coder |
| Moonshot (Kimi) | kimi-k2.5, kimi-k2 |
| Zhipu GLM | glm-5, glm-4.7-flash, glm-4.6v |
| Baidu ERNIE | ernie-4.5-thinking, ernie-4.5-vl |
| ByteDance Seed | seed-2.0-lite, seed-2.0-mini |
Capability Mapping
The skill uses keyword-based pattern matching to infer model capabilities:
| Keyword | Capability | Primary Use Case |
|---|---|---|
vl, vision, image, 4v, 4o |
Vision/Multimodal | Image analysis, OCR, chart interpretation |
code, coder, codex |
Code Generation | Programming, debugging, refactoring |
o1, o3, o4, reasoning, thinking |
Advanced Reasoning | Mathematical reasoning, complex logic |
max, pro, premium, large, opus |
Premium/Tier-1 | High-quality output, complex tasks |
mini, small, lite, flash, haiku, nano |
Lightweight | Fast response, simple tasks, cost-sensitive |
Recommendation Logic
1. Parse ~/.openclaw/openclaw.json to extract configured models
2. Analyze user's task requirements
3. Match task to model capabilities via keyword detection
4. Return the best-matching model from available configuration
5. Fall back to default model if no specific match found
Usage Examples
User: "Which model should I use for coding?" → Recommend: minimax-m2.5, deepseek-coder, qwen3-coder-plus
User: "I need to analyze an image" → Recommend: qwen3-vl, glm-4.6v, gpt-4o, gemini-1.5-pro
User: "For complex reasoning tasks" → Recommend: o3, claude-opus-4.6, qwen3-max, gemini-3.1-pro
Notes
- Automatically adapts to your specific configuration
- Prioritizes models that are actually available in your setup
- Falls back gracefully when exact capability matches aren't found
- Supports both global (OpenAI, Claude, Gemini) and domestic (Chinese) model providers
Usage Guidance
This skill appears to do what it says, but before enabling it: 1) inspect your ~/.openclaw/openclaw.json to confirm it contains only model IDs/metadata and not API keys or secrets; 2) note that the skill's metadata does not declare the config path it will read — prefer skills that explicitly declare required config access; 3) if you are concerned about data exfiltration, restrict network access for the agent or run the skill in a sandbox until you've verified behavior; and 4) if you need stronger guarantees, ask the author to add the required-config-path to the metadata and to limit parsing to non-sensitive fields.
Capability Analysis
Type: OpenClaw Skill
Name: local-config-model-recommender
Version: 1.0.0
The skill instructs the AI agent to read and parse the user's local configuration file (~/.openclaw/openclaw.json) in SKILL.md. This file is highly sensitive as it typically contains API keys and provider credentials for various AI services. While this access is plausibly needed to determine which models are available for recommendation, the broad instruction to read the entire config file poses a significant risk of exposing secrets within the agent's context. No explicit evidence of data exfiltration or malicious intent was found, but the behavior is high-risk.
Capability Assessment
Purpose & Capability
The skill's name and description align with its runtime instructions (it reads the OpenClaw config and recommends models). However, the package metadata declares no required config paths even though SKILL.md explicitly instructs the agent to parse ~/.openclaw/openclaw.json — a mismatch in declared vs. actual resource access.
Instruction Scope
Instructions are narrowly scoped: parse the OpenClaw config, analyze task keywords, and match model IDs. They do not instruct network exfiltration or broad system reads. One minor concern: the guidance to 'analyze user's task requirements' is vague and could allow broader context gathering unless the agent's runtime policies restrict that.
Install Mechanism
This is an instruction-only skill with no install spec or code files, so nothing is written to disk and there is no external install risk.
Credentials
The skill declares no environment variables or credentials, which is reasonable. But it reads ~/.openclaw/openclaw.json; depending on your setup that file could contain more than model IDs (e.g., endpoint URLs or API keys). The skill does not declare this config access in its metadata, so verify what is stored in that file before allowing the skill to read it.
Persistence & Privilege
The skill is not always-enabled and does not request elevated or persistent privileges in the metadata. Autonomous invocation is allowed by default (platform normal) and is not a standalone red flag here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install local-config-model-recommender - After installation, invoke the skill by name or use
/local-config-model-recommender - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of local-config-model-recommender.
- Dynamically reads user’s OpenCLAW configuration to list available models.
- Recommends optimal AI models based on keyword pattern matching aligned to task requirements.
- Supports a wide range of both global and domestic models including OpenAI, Claude, Gemini, DeepSeek, Kimi, Zhipu GLM, Qwen, and MiniMax.
- Covers multimodal, code generation, advanced reasoning, premium tier, and lightweight models for various needs.
- Ensures recommendations are tailored to models actually available in your configuration.
Metadata
Frequently Asked Questions
What is local-config-model-recommender?
Intelligently recommends optimal AI models based on task requirements. Dynamically reads the user's OpenCLAW configuration and provides context-aware model s... It is an AI Agent Skill for Claude Code / OpenClaw, with 265 downloads so far.
How do I install local-config-model-recommender?
Run "/install local-config-model-recommender" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is local-config-model-recommender free?
Yes, local-config-model-recommender is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does local-config-model-recommender support?
local-config-model-recommender is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created local-config-model-recommender?
It is built and maintained by garygou1024 (@garygou1024); the current version is v1.0.0.
More Skills