Skill Claude CodeCodex
Run multi-round AI brainstorming debates between multiple LLM providers (GPT, DeepSeek, Groq, Ollama). Claude actively participates as a debater alongside external models. Use when the user wants diverse perspectives, multi-model critiques, or synthesized answers from several AI models working together.
Agent
Execute brainstorm prompts using Gemini.
Agent
Execute brainstorm prompts using GPT.
Agent
Execute brainstorm prompts using Claude Haiku.
Agent
Execute brainstorm prompts using Claude Opus.
Agent
Execute brainstorm prompts using Claude Sonnet.
Agent
Orchestrate multi-model brainstorming debates using models available in your environment. No API keys needed.
MCP server Claude CodeCodexCursor +2
Multi-model AI brainstorming debates with code review and quick perspectives. Runs locally from the brainstorm-mcp npm package. Needs 3 environment variables to run.
Skill Claude CodeCodex
Track projects, epics, tasks, subtasks, notes, and decisions using the saga-mcp project tracker. Use when the user asks to track work, manage tasks, plan features, log decisions, or when you need persistent context across coding sessions.
MCP server Claude CodeCodexCursor +2
Jira-like project tracker for AI agents with SQLite, full task hierarchy, and dashboard. Runs locally from the saga-mcp npm package.
Hook Claude Code
Runs when you submit a prompt, before the agent sees it, running SWARMCODE_REDIS_URL=redis://192.168.4.161:30379 SWARMCODE_WO. From spranab/swarmcode.
Settings file Claude Code
Agent settings declaring 1 hook event (UserPromptSubmit).
MCP server Claude CodeCodexCursor +2
Real-time communication between Claude Code instances across machines and workspaces. Redis-backed channels with background task notifications for true push in VS Code. Runs locally from the mcp-agent-bridge npm package. Needs 2 environment variables to run.
Instructions file
Instructions for spranab/swarmcode, covering swarmcode, architecture, key design, running and environment.