CLAUDE.md · diff

git:20261011.a907070 to git:20261012.0f01137

2 added, 2 removed. Audit A to A.

# UniFi MCP Server - Claude Instructions
This file provides project-specific instructions for AI coding assistants working on the UniFi MCP Server.
## Project Overview
The UniFi MCP Server is a Model Context Protocol (MCP) server that exposes the UniFi Network Controller API, enabling AI agents and applications to interact with UniFi network infrastructure in a standardized way.
- **Current Version**: v0.5.0 (package version, tag pending; last tagged release is v0.2.5; see `DEVELOPMENT_PLAN.md` §5 for release-versioning notes)
+ **Current Version**: v0.6.0 (see `DEVELOPMENT_PLAN.md` §5 for release-versioning notes)
**Python Version**: 3.10+
**Framework**: FastMCP
## Quick Start for AI Assistants
### Before Starting Work
1. **Read Key Documentation**:
- `README.md` - Project overview and features
- `AGENTS.md` - Universal AI agent guidelines
- `DEVELOPMENT_PLAN.md` - Roadmap and priorities
- `TODO.md` - Current tasks and phase breakdown
- `API.md` - Complete MCP tool documentation
2. **Understand the Architecture**:
- `src/main.py` - MCP server entry point
- `src/api/` - UniFi API client
- `src/models/` - Pydantic data models
- `src/tools/` - MCP tool implementations
- `tests/unit/` - Unit tests (3,152 tests passing, 93% line coverage)
### Development Workflow
1. **Feature Development**:
- Create feature branch: `git checkout -b feature/your-feature`
- Follow TDD: Write tests first, then implementation
- Maintain 80% minimum test coverage for new code
- Use Pydantic models for all data structures
- Add comprehensive docstrings (Google style)
2. **Code Quality**:
- Run tests: `pytest tests/unit/`
- Format: `black src/ tests/` and `isort src/ tests/`
- Lint: `ruff check src/ tests/ --fix`
- Type check: `mypy src/`
- Pre-commit: `pre-commit run --all-files`
- Or via `make`: `make lint`, `make test`, `make typecheck`, and `make ci` for every blocking CI check (`make help` lists all targets)
3. **Safety Mechanisms**:
- All mutating operations require `confirm=True`
- Implement dry-run mode for preview
- Add audit logging for operations
- Validate all user inputs
- Never commit secrets or credentials
### Technology Stack
- **Language**: Python 3.10+
- **Framework**: FastMCP (MCP server framework)
- **API Client**: httpx (async HTTP)
- **Data Validation**: Pydantic v2
- **Testing**: pytest with asyncio support
- **Caching**: Redis (optional)
- **Monitoring**: agnost.ai (optional)
### API Modes
The server supports three UniFi API access modes:
1. **Local Gateway API** (Recommended): Full feature support
- `UNIFI_API_TYPE=local`
- `UNIFI_LOCAL_HOST=192.168.2.1`
2. **Cloud V1 API**: Stable, aggregate statistics only
- `UNIFI_API_TYPE=cloud-v1`
3. **Cloud EA API**: Early Access, aggregate statistics only
- `UNIFI_API_TYPE=cloud-ea`
### Current Development Focus
Phases 0–3, 5a (Protect v7) and 6 (Mobility, InnerSpace, Carrier Fabric) are complete. **Phase 4** (testing, polish, minor gaps, developer experience) is active, and Phase 5 is partially started. See `TODO.md` and `DEVELOPMENT_PLAN.md` for the live status of each item.
- **Phase 4 open items**: release tag (v0.5.0, pending hardware verification).
+ **Phase 4 open items**: none (v0.6.0 released; v0.5.0 was tagged but not published, see `DEVELOPMENT_PLAN.md` §5).
**Phase 5 already in place**: A2A agent card and endpoints (behind bearer auth), `UNIFI_PROFILE` tool profiles, `UNIFI_READ_ONLY` and `UNIFI_DRY_RUN` modes, bearer auth on network transports, roles over risk tiers with server-issued tokens, hash-chained wrapper-level audit log, `/metrics`, and the fleet stack (Phases 0–4 of `docs/FLEET_SCALING_PLAN.md`: routing, Postgres registry, pool and read cache, several replicas, workers, `fleet_*` read tools).
**Total**: 343 MCP tools registered in local mode across 60 tool modules, 3,152 unit tests passing
### Important Constraints
1. **UniFi Network 9.0+ Required**: Some features require Network 9.0+
2. **Local API Recommended**: Cloud APIs have limited functionality
3. **Endpoint Verification**: Some documented API endpoints may not exist in all versions
4. **Testing**: Integration tests require real UniFi hardware
### Getting Help
- **Issues**: [GitHub Issues](https://github.com/enuno/unifi-mcp-server/issues)
- **Documentation**: See `API.md` for complete tool reference
- **Examples**: Check `docs/examples/` for AI assistant prompts
## Key Principles
1. **Safety First**: Never perform destructive operations without confirmation
2. **Quality Over Speed**: Maintain high test coverage and code quality
3. **Clarity**: Write self-documenting code with clear docstrings
4. **Consistency**: Follow existing patterns and conventions
5. **Security**: Never commit credentials, validate all inputs
## Additional Resources
- [CONTRIBUTING.md](CONTRIBUTING.md) - Contribution guidelines
- [SECURITY.md](SECURITY.md) - Security policy and best practices
- [AGENTS.md](AGENTS.md) - Detailed AI agent guidelines
- [TESTING_PLAN.md](docs/archive/TESTING_PLAN.md) - Testing strategy
- [DEVELOPMENT_PLAN.md](DEVELOPMENT_PLAN.md) - Complete roadmap
---
**Last Updated**: 2026-10-10
**Maintained By**: Development Team