darrenhinde/OpenAgentsControl — reverse-engineered prompt

Build me an OpenCode based AI agent toolkit that helps me get consistent, production ready code instead of generic AI output.

I want it to learn my team’s coding patterns from simple editable context files, then use those patterns whenever it writes code. The flow should be plan first, show me the plan, wait for approval, then execute step by step with validation, testing, and code review built in. It should feel safe and controlled, not like the agent is running off on its own.

Please make it work across common languages like TypeScript, Python, Go, and Rust, and keep it model agnostic so I can use different AI providers or local models. I also want specialized helper agents for things like finding context, writing code, reviewing code, and checking tests.

Make setup easy with a one command install, an update path, and clear docs. If it makes sense, include a Claude Code plugin option too. Look up current OpenCode docs online if you need to.