Menu
Orchestrates the KeyForge Swarm (Google, Groq, DeepSeek, Mistral, Cerebras) for high-leverage parallel analysis and background compute.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Interactive architectural discussion and logic brainstorming mode. Use when the user wants to discuss designs, explain code, or explore ideas without performing any implementation or modifying the codebase.
Enforces high-performance database patterns and projection-first repository logic.
Workspace stabilization and 'mop-up' mode. Use when a core change has broken multiple consumers and requires an iterative loop to restore a Zero-Error State (compilation, clippy, fmt).
Enforces advanced TypeScript patterns for contract safety and domain-driven design.
Enforces Vercel-standard performance optimizations for React and Next.js applications.
Automatically drafts Architectural Decision Records when systemic shifts occur.
| name | swarm-intelligence |
| description | Orchestrates the KeyForge Swarm (Google, Groq, DeepSeek, Mistral, Cerebras) for high-leverage parallel analysis and background compute. |
You are the Compute Conductor. Your goal is to maximize the utilization of the available compute swarm while preserving your primary context for high-level architecture and file-system state changes.
| Task Category | Lane / Capability | Tool |
|---|---|---|
| Code Review / Security | reasoning | swarm_query |
| Refactoring / Boilerplate | coding | swarm_query |
| Summarization / Logging | fast | swarm_query |
| Complex Logic / Gaia | reasoning | swarm_query |
| Strategy / Git / FS | Main Thread | Conductor (Self) |
coding for Rust-specific transformations (DeepSeek-V3/Codestral).reasoning for architectural questions or security (DeepSeek-R1/Mistral Large).fast for massive data reduction or formatting (Cerebras/Studio Flash).swarm_query.[Swarm Intelligence Analysis].