Brummer MCP Server

# External Context Sources: Agentic AI Coders ## Primary Documentation ### BubbleTea Framework - **Documentation**: [BubbleTea Guide](https://github.com/charmbracelet/bubbletea/tree/master/tutorials) - Component composition and view management - **Key Insights**: Model-View-Update pattern, command handling, subscription patterns - **Implementation Guidance**: New AI coder view follows existing view patterns, async operations via Cmd ### Model Context Protocol (MCP) - **Specification**: [MCP 2025-06-18](https://modelcontextprotocol.io/specification/2025-06-18) - Protocol for AI tool integration - **Key Sections**: Tool registration, resource management, client-server communication - **Implementation Focus**: New AI coder tools, session management, streaming capabilities ### AI Integration Patterns - **OpenAI Go SDK**: [OpenAI Go](https://github.com/sashabaranov/go-openai) - Go client for AI API integration - **Claude API**: [Anthropic API](https://docs.anthropic.com/en/api/getting-started) - Claude integration patterns - **Key Patterns**: Streaming responses, context management, error handling, rate limiting ### Go Concurrency Patterns - **Go Blog**: [Concurrency Patterns](https://blog.golang.org/pipelines) - Pipeline and worker pool patterns - **Relevance**: AI coder task management, async processing, resource coordination - **Application**: Background AI tasks, status monitoring, event processing ## Industry Standards & Best Practices ### Security Standards - **OWASP Go Security**: [Go Security Guide](https://owasp.org/www-project-go-secure-coding-practices-guide/) - Secure coding practices - **Applicable Guidelines**: Input validation, sandbox isolation, credential management - **Threat Model**: AI coder workspace isolation, API key protection, file system security ### Performance Standards - **Go Performance**: [Go Performance Tips](https://github.com/golang/go/wiki/Performance) - Memory management and optimization - **Target Metrics**: <100ms TUI response time, <500MB memory per AI coder instance - **Optimization**: Connection pooling, context reuse, efficient event handling ### API Design Standards - **REST Guidelines**: [Go API Standards](https://github.com/golang-standards/project-layout) - Project structure and API design - **MCP Compliance**: Tool naming conventions, error handling, resource management - **Implementation**: Consistent tool interfaces, structured error responses ## Reference Implementations ### Similar AI Integration Projects - **Continue.dev**: [Continue VSCode Extension](https://github.com/continuedev/continue) - AI coding assistant integration - **Pattern Demonstrated**: AI coder lifecycle, workspace management, streaming UI updates - **Adaptation Needed**: Terminal UI instead of VSCode, MCP integration, Go implementation ### Process Management Examples - **Docker CLI**: [Docker CLI Source](https://github.com/docker/cli) - Container lifecycle management - **Integration Pattern**: Process spawning, status monitoring, resource isolation - **Lessons Learned**: Graceful shutdown, resource cleanup, status aggregation ### Terminal UI References - **Lazygit**: [Lazygit TUI](https://github.com/jesseduffield/lazygit) - Complex TUI with multiple views - **Architecture Pattern**: View switching, keyboard navigation, async operations - **Trade-offs**: Performance vs features, keyboard shortcuts, visual hierarchy ### Event-Driven Architecture - **NATS Messaging**: [NATS Go Client](https://github.com/nats-io/nats.go) - Pub/sub messaging patterns - **System Design**: Event sourcing, async processing, error handling - **Scalability Patterns**: Worker pools, backpressure, graceful degradation ## Standards Applied ### Coding Standards - **Go Style Guide**: [Effective Go](https://golang.org/doc/effective_go.html) - Naming, error handling, concurrency - **Specific Rules**: - Package naming: `ai_coder` or `aicoder` (avoid mixed case) - Interface naming: `AICoderManager`, `WorkspaceIsolator` - Error handling: Explicit error returns, context propagation ### API Design Standards - **MCP Tool Naming**: [MCP Tool Guidelines](https://modelcontextprotocol.io/specification/2025-06-18#tools) - Consistent naming conventions - **Naming Patterns**: `ai_coder_create`, `ai_coder_status`, `ai_coder_workspace` - **Error Handling**: Structured error responses with codes and messages ### Testing Standards - **Go Testing**: [Go Testing Guide](https://golang.org/doc/tutorial/add-a-test) - Unit and integration testing - **Coverage Requirements**: >80% coverage for AI coder core logic - **Test Patterns**: Table-driven tests, mock AI providers, integration test isolation ### Security Standards - **Credential Management**: Environment variables, no hardcoded secrets - **Workspace Isolation**: Chroot/container-like restrictions for AI coder file access - **API Security**: Rate limiting, input validation, secure credential storage ## External Dependencies Assessment ### AI Provider APIs - **OpenAI API**: Rate limits, pricing considerations, streaming support - **Claude API**: Context window management, tool use capabilities - **Local Models**: Ollama integration, resource requirements, performance trade-offs ### System Dependencies - **File System**: Workspace sandboxing requirements (potentially containers/chroot) - **Process Management**: Resource limits, cgroup integration - **Network**: Outbound API access, proxy support, timeout handling ### Development Tools - **Testing Frameworks**: Testify for assertions, httptest for API mocking - **Build Tools**: Go modules, cross-platform compilation - **Documentation**: godoc integration, README maintenance ## Implementation Decision Framework ### AI Provider Selection - **Criteria**: API stability, Go SDK quality, feature completeness, cost - **Default Choice**: Multiple provider support with plugin architecture - **Configuration**: Provider selection via configuration files ### Workspace Isolation Strategy - **Options**: Directory restrictions, containers, virtual filesystems - **Recommendation**: Start with directory restrictions, evolve to containers - **Trade-offs**: Security vs complexity, performance vs isolation ### UI Architecture Approach - **Pattern**: Follow existing BubbleTea view pattern from TUI - **Components**: Reuse existing list, viewport, textinput components - **Extensions**: Custom AI coder status components, progress indicators ### Error Handling Strategy - **Approach**: Structured errors with context, graceful degradation - **User Experience**: Clear error messages, recovery suggestions - **Logging**: Comprehensive logging for debugging and monitoring