---
description: Get multi-model AI perspectives on any coding problem
argument-hint: Your question or problem
---
# /polydev-ask [question] - Get Multi-Model AI Perspectives
Query multiple AI models (GPT-5, Gemini, Grok, GLM) simultaneously to get diverse perspectives on any coding problem.
## Instructions
When the user runs `/polydev-ask [question]`, do the following:
1. **Extract the question**: Parse the argument after `/polydev-ask` as the question. If no question is provided, look at the recent conversation context.
2. **Enhance the question**: Add relevant context from:
- Current file being worked on
- Recent errors or debugging attempts
- Project structure if relevant
3. **Check if MCP tools are available** by looking for `mcp__plugin_polydev_mcp__get_perspectives` in your tools.
### If MCP tools ARE available:
Call the perspectives tool:
```
mcp__plugin_polydev_mcp__get_perspectives({
prompt: "The enhanced question with full context"
})
```
### If MCP tools are NOT available (Cowork / sandboxed environments):
Use curl to call the Polydev API directly:
```bash
# Load token if not in env
[ -z "$POLYDEV_USER_TOKEN" ] && [ -f "$HOME/.polydev.env" ] && source "$HOME/.polydev.env"
if [ -z "$POLYDEV_USER_TOKEN" ]; then
echo "Not authenticated. Run /polydev:login first."
exit 1
fi
curl -s -X POST "https://www.polydev.ai/api/mcp" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $POLYDEV_USER_TOKEN" \
-H "User-Agent: polydev-mcp/1.9.0" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "get_perspectives",
"arguments": {
"prompt": "YOUR ENHANCED QUESTION HERE"
}
},
"id": "1"
}'
```
4. **Synthesize the response**:
- Highlight areas of **consensus** (high confidence)
- Note where models **differ** (needs consideration)
- Provide actionable **recommendations**
## Example Usage
**User**: `/polydev-ask How should I structure my React state management?`
**Claude calls** (if MCP tools available):
```
mcp__plugin_polydev_mcp__get_perspectives({
prompt: "How should I structure state management in a React application? I'm building a dashboard with real-time updates, user preferences, and form state."
})
```
**Claude synthesizes**:
> **Multi-Model Perspectives on React State Management**
>
> **Consensus (High Confidence)**:
> - All models recommend separating server state (React Query/SWR) from client state
> - Zustand is favored for simplicity in medium-sized apps
>
> **Different Perspectives**:
> - GPT-5 emphasizes Redux Toolkit for large teams
> - Gemini recommends Jotai for atomic state
> - Grok suggests starting simple with useState + Context
>
> **Recommendation**: Start with Zustand for client state + React Query for server state.
## When to Use
- **Stuck debugging**: After 2-3 failed attempts
- **Architecture decisions**: Choosing between approaches
- **Code review**: Getting validation on implementation
- **Best practices**: Unsure about patterns
## Models Queried
All queries consult 4+ models in parallel:
- **GLM-4.7** - Zhipu AI's flagship model
- **Gemini 3** - Google's reasoning model
- **Grok 4.1** - xAI's inference model
- **GPT-5** - OpenAI's model
