Agent Switchboard

Route this local codebase error analysis task to Claude, Codex, and Gemini: have each model identify the root cause, run a cross-model debate, then produce the most reliable fix plan while reusing shared context to minimize token usage.

Returns each model’s diagnosis, key disagreements, and a consolidated best fix recommendation.

Set up an agent routing workflow for the current project: assign refactoring to the model best at cleanup, test generation to the model best at testing, and change explanations to the model best at documentation, all coordinated through the local MCP server.

Produces a clear routing plan along with code changes, tests, and explanatory notes.

Share the core context of this repository across multiple AI coding agents, syncing only necessary information instead of resending the full codebase; then have different models handle requirement breakdown, implementation suggestions, and risk assessment.

Returns an efficient shared-context setup plus model-specific planning, implementation advice, and risk analysis.