SuperAgent & Sequential MCP
Relevant source files
The following files were used as context for generating this wiki page:
SuperAgent and Sequential MCP represent the core orchestration and reasoning components of the SuperGemini Framework. These Model Context Protocol (MCP) integrations extend the standard capabilities of the Gemini CLI, enabling multi-agent orchestration and structured multi-step reasoning within a unified development workflow. SuperAgent allows for sub-agent orchestration similar to the SuperClaude ecosystem, while Sequential MCP facilitates complex project analysis and systematic debugging.
Together, these modules form an enhancement layer that transforms the Gemini CLI into a sophisticated development environment capable of handling complex, multi-domain software engineering tasks. They are integrated through protocol abstraction and health monitoring, ensuring seamless coordination between the primary AI and external specialized tools.
Sources: Docs/Developer-Guide/README.md:46-56, CLAUDE.md:475-485, CHANGELOG.md:28-35
SuperAgent MCP Integration
The SuperAgent MCP integration, specifically registered as superagent:gemini, allows the framework to perform sub-agent orchestration. This enables the framework to match multi-agent workflows found in related frameworks like SuperClaude. It operates by coordinating between different specialized AI specialists to handle complex tasks that require diverse domain expertise.
Sub-agent Coordination
SuperAgent facilitates the interaction between 13 specialized domain experts within the framework. When a command is issued, SuperAgent helps manage the transition and data flow between these personas.
| Agent Persona | Role |
|---|---|
| system-architect | High-level system design and structural planning |
| security-engineer | Security analysis and veto authority |
| quality-engineer | Testing, validation, and quality standards |
| technical-writer | Documentation generation and maintenance |
Sources: CLAUDE.md:154-175, CHANGELOG.md:28-33, Docs/Getting-Started/quick-start.md:38-44, SuperGemini/Commands/index.toml:6-8
The diagram above shows how the SuperAgent MCP acts as a coordinator between the core command parser and specialized domain agents to produce a unified response.
Sources: Docs/Getting-Started/quick-start.md:52-60, CHANGELOG.md:28-33
Sequential MCP Operations
The Sequential MCP is designed for multi-step reasoning engines. It is a critical component for commands that require systematic analysis, such as project indexing, debugging, and complex system design. It allows the AI to break down a large problem into a series of logical steps, maintaining state and context throughout the process.
Multi-step Reasoning Workflow
The Sequential MCP is typically activated via the --seq flag or automatically triggered by high-complexity commands like /sg:index or /sg:troubleshoot.
- Systematic Analysis: Breaking down the codebase or problem into components.
- Step-by-Step Execution: Processing each component sequentially to avoid context loss.
- Synthesis: Combining findings from each step into a final solution or report.
Sources: CLAUDE.md:478, Docs/Reference/README.md:104-106, SuperGemini/Commands/index.toml:19-25
The sequence diagram illustrates the iterative nature of the Sequential MCP when performing systematic project analysis.
Sources: SuperGemini/Commands/index.toml:34-40, Docs/Reference/README.md:104-106
Integration and Configuration
Both SuperAgent and Sequential MCP are managed by the framework's installation engine. They are configured via JSON files and landed in the active mcpServers section of the Gemini configuration.
Configuration Parameters
Configuration for these servers includes the command to execute the server and any necessary environment variables.
| MCP Server | Config File | Primary Function |
|---|---|---|
superagent:gemini | superagent.json | Sub-agent orchestration and multi-agent workflows |
sequential | sequential.json | Multi-step reasoning and systematic analysis |
Sources: CLAUDE.md:475-485, CHANGELOG.md:28-35
Installation and Health Monitoring
The framework ensures these servers are properly registered within the ~/.gemini/ directory. During the installation process, the system validates prerequisites such as Node.js and npm, which are required for most MCP integrations.
# To install or update MCP components
SuperGemini install --components mcp --force
Sources: CLAUDE.md:105-115, CLAUDE.md:488-495
Conclusion
SuperAgent and Sequential MCP are fundamental to the SuperGemini Framework's ability to provide "Structured Development Capabilities." SuperAgent provides the horizontal coordination between domain experts, while Sequential MCP provides the vertical depth required for complex multi-step reasoning. These components ensure that the framework can scale from simple queries to full-scale architectural design and project management.
Sources: README.md:65-75, Docs/Getting-Started/quick-start.md:78-85