render

Overview

Runs the external codex CLI in a tight loop: first to execute your prompt, then immediately to critique the result and propose next actions. Continue iterating by providing new prompts until you finish with /done.

Usage

ouro render [--codex-model <model>] [prompt]

Key Options

• --codex-model <model> – Override the Codex model used for both execution and validation (default gpt-5).
• prompt – Optional positional text to seed the first Codex run; otherwise the command prompts interactively.

Flow

1. Ouro streams the raw Codex execution output.
2. It builds a validation prompt that summarises the user request and Codex response.
3. Codex analyses the result, listing issues, missing items, and suggested next actions.
4. You can enter another instruction, using the validation feedback as guidance, or type /done to exit.

When to Use

• You want a structured Codex dialogue that alternates between implementation and self-review.
• You need a light-weight alternative to ouro x while still keeping validation pressure on each Codex run.

Follow-up Checklist

• Apply or discard Codex changes based on the validation notes.
• Run project tests or linters suggested by the validation step before committing.