AI Coding Agent for Existing Code Optimization

I’m looking for a senior AI coding-agent / software engineering specialist to help set up an AI-assisted development workflow for our existing codebase. This is an existing, actively developed codebase. What I need Set up and optimize OpenCode, OpenRouter, Command Code Create specialized agents for: * Code Review & Audit * Coding / Feature Development * Bug Fixing * Refactoring & Optimization * Frontend Development * Backend Development * Testing * Documentation * Create strong project skills, rules, instructions and context * Review our existing `AGENTS.md`, `ARCHITECTURE.md`, TODO files and other documentation and build the new MD files **to match our existing architecture, coding standards, documentation style and project structure * Ensure agents inspect existing code and dependencies before making changes and avoid unnecessary/destructive modifications * Set up a proper Git workflow with reviewable changes and meaningful commits * Agents should be capable of writing, updating, refactoring, optimizing and safely removing code * Support both backend and frontend changes while following the existing architecture and UI/design system Deliverables By the end of the project, I expect: 1. Working Ai Agent setup & configured for our repository. 2. Model/provider configuration with recommendations based on quality, speed and cost. 3. Multiple task-specific agents configured and tested. 4. A structured skills system for the different development workflows. 5. A properly structured `AGENTS.md` and supporting `.md` files based on our existing documentation — not generic templates. 6. Clear rules for architecture, coding standards, security, database, Git and destructive changes. 7. Dedicated workflows for audit → review → implementation → testing → commit. 8. Git-aware workflow where agents produce clean, focused and meaningful commits. 9. A working demonstration on the actual existing codebase — including code review/audit and a real code change. 10. Short documentation explaining how I can use each agent and when to use it. Our stack includes React, TypeScript, Node.js, Express, PostgreSQL, BullMQ, Redis/Dragonfly, Zustand, Tailwind, NextUI, Zod etc. The goal is to have a reliable AI development environment where I can choose the appropriate agent for review, audit, coding, optimization, fixing, testing or frontend/backend development without risking the existing codebase.