# opencode Configuration

This file configures opencode for working with the DataPRO codebase.

## Model Configuration

Configured to use vast.ai hosted GLM-5-FP8 model with 131k context window.

## MCP Servers

### vastai-ctl
Control vast.ai GPU instances (start, stop, status).

### retrieval
Semantic search over code and enrichment documentation.
- `vector_search(query)` - Find relevant code by meaning
- `get_module_summary(path)` - Get AI-generated module docs
- `list_enriched_modules()` - List all enriched modules

## AI-Generated Artifacts

This codebase has pre-computed AI artifacts for efficient assistance:

| Location | Purpose |
|----------|---------|
| `GLM5Analysis/Architecture.md` | System architecture overview |
| `GLM5Analysis/PatternLibrary/` | Reusable code patterns |
| `GLM5Analysis/PromptTemplates/` | Task-specific guides |
| `GLM5Analysis/TestScaffolds/` | Test templates |
| `enriched-qwen3-coder-next/` | Module-level documentation |

**Always check `GLM5Analysis/` before starting a task.**

## Workflow

1. Check `GLM5Analysis/Architecture.md` for system context
2. Check `GLM5Analysis/PromptTemplates/` for task-specific guidance
3. Use `vector_search()` to find relevant code
4. Use `get_module_summary()` for detailed module docs
5. Follow patterns from `GLM5Analysis/PatternLibrary/`

## Build Commands

```bash
# Build solution (requires Windows/Visual Studio)
msbuild DataPRO/DataPRO.sln /p:Configuration=Debug

# Run tests
# Tests are NUnit-based in DataPRO/UnitTest/
```

## Important Files

- `GIT Migration.md` - Plan for SVN to Git migration
- `AGENTS.md` - Instructions for AI assistants