feat: init progettino - OpenCode environment setup

- AGENTS.md with agent instructions
- README with project overview
- .gitignore for the project structure

.gitignore

@@ -0,0 +1,17 @@
+# OpenCode
+.opencode/node_modules/
+.opencode/package-lock.json
+.opencode/package.json
+.opencode/.gitignore
+.opencode/memory/.gitignore
+
+# Supertonic (separate repo)
+/supertonic/
+
+# Firecrawl scraped data
+.firecrawl/
+
+# Python
+__pycache__/
+*.pyc
+.venv/

AGENTS.md

@@ -0,0 +1,173 @@
+# OpenCode Agent Instructions
+
+## Overview
+
+This repository contains an OpenCode development environment setup. OpenCode is an AI-powered development assistant that helps you build, test, and deploy software. The environment is configured with the OpenCode plugin and runs in a Linux-based development container.
+
+## Project Context
+
+- **Project Type**: AI Agent Development Environment
+- **Main Technology**: Node.js with OpenCode plugin
+- **Target Audience**: AI agents and developers working on code generation projects
+- **Development Server**: Local development with Ollama for on-premise LLMs
+
+## Development Environment
+
+### System Setup
+
+- **OS**: Debian Linux (running in Proxmox container)
+- **IP Address**: 192.168.0.240
+- **Memory**: Standard Linux environment with Node.js
+- **AI Models**: Ollama server at http://192.168.0.50:11434 (gemma2:9b, nomic-embed-text)
+
+### Key Directories
+
+- `.opencode/` - OpenCode plugin and configuration
+- `.opencode/memory/` - Persistent memory storage
+- `.opencode/node_modules/` - Node.js dependencies
+
+## OpenCode Usage
+
+### Basic Commands
+
+```bash
+# Start OpenCode agent
+opencode
+
+# Run with specific tool
+opencode --tool claude
+
+# Check OpenCode version
+opencode --version
+```
+
+### Agent Workflow
+
+1. **Initialization**: OpenCode loads the project context and available skills
+2. **Task Analysis**: The agent analyzes the request and breaks it down into steps
+3. **Execution**: Tasks are delegated to specialized subagents
+4. **Verification**: Results are checked against requirements
+5. **Documentation**: Progress is saved to memory and AGENTS.md
+
+## Important Development Commands
+
+### Node.js Management
+
+```bash
+# Install dependencies
+npm install
+
+# Run scripts (if defined in package.json)
+npm run <script-name>
+
+# Check for outdated packages
+npm outdated
+
+# Update packages
+npm update
+```
+
+### Git Operations
+
+```bash
+# Check status
+git status
+
+# Add changes
+git add .
+
+# Commit changes
+git commit -m "message"
+
+# Push to remote
+git push
+```
+
+### File Operations
+
+```bash
+# Search for files
+find . -name "*.js" -type f
+
+# Search content
+grep -r "pattern" .
+
+# List all files
+ls -la
+```
+
+## Testing and Debugging
+
+### Diagnostic Tools
+
+- **lsp_diagnostics**: Check for errors and warnings in files
+- **grep**: Search for patterns across the codebase
+- **ast_grep**: Advanced code pattern matching
+- **read**: Read file contents
+
+### Debugging Strategy
+
+1. Use `lsp_diagnostics` to identify issues
+2. Search for similar patterns with `grep` or `ast_grep`
+3. Consult documentation for external libraries
+4. Test changes in isolation
+5. Verify with build/test commands
+
+## Conventions and Best Practices
+
+### Code Quality
+
+- **Type Safety**: Never suppress type errors with `as any` or `@ts-ignore`
+- **Error Handling**: Always handle errors properly; avoid empty catch blocks
+- **Testing**: Write tests for new functionality; don't delete failing tests
+- **Refactoring**: Refactor safely; use tools to ensure correctness
+
+### Workflow
+
+- **Atomic Commits**: Each commit should represent a single logical change
+- **PRs**: Keep pull requests focused and small
+- **Documentation**: Update documentation as part of any significant change
+- **Memory**: Use memory files to persist important information
+
+## Common Gotchas
+
+### Environment Issues
+
+- **Node.js Version**: Ensure you're using a compatible Node.js version
+- **Dependencies**: Install all dependencies before running commands
+- **Permissions**: Check file permissions for read/write access
+
+### OpenCode Specific
+
+- **Context Window**: Keep requests focused to avoid context overflow
+- **Skill Selection**: Use appropriate skills for the task domain
+- **Background Tasks**: Use `run_in_background=true` for parallel operations
+- **Task Continuation**: Use `task_id` to continue previous agent sessions
+
+## External Resources
+
+### Documentation Links
+
+- [OpenCode Documentation](https://opencode.ai/docs)
+- [Node.js Documentation](https://nodejs.org/docs)
+- [Ollama Documentation](https://ollama.com/docs)
+
+### Local Servers
+
+- **Ollama**: http://192.168.0.50:11434
+- **Development Server**: (if applicable)
+
+## Project-Specific Notes
+
+This is a minimal development environment primarily used for:
+- Testing OpenCode agent capabilities
+- Developing AI-assisted coding workflows
+- Experimenting with on-premise LLM integration
+
+The repository is intentionally simple to serve as a clean slate for various development tasks.
+
+## Updating This File
+
+Keep AGENTS.md up to date with any new conventions, commands, or structural changes. Remove outdated information and add new insights as they're discovered.
+
+(End of file)

README.md

@@ -0,0 +1,19 @@
+# Progettino
+
+Ambiente di sviluppo **OpenCode** — un harness per agenti AI basato su OpenCode plugin.
+
+## Contenuto
+
+- `AGENTS.md` — Istruzioni per l'agente OpenCode (convenzioni, comandi, ambiente)
+- `.opencode/` — Plugin e configurazione OpenCode
+
+## Ambiente
+
+- **OS**: Debian 13 (container Proxmox)
+- **IP**: 192.168.0.240
+- **LLM locale**: Ollama su 192.168.0.50:11434 (gemma2:9b)
+- **OpenCode**: Plugin v1.14.48
+
+## Vedi anche
+
+- [Supertonic 3 TTS Demo](http://192.168.0.183:8085/truetype/supertonic-local-demo) — TTS locale 31 lingue