# Review on Codex

Claude Code builds each Ticket and Codex reviews it, by giving the review Step its own harness and model.

To have Codex review what Claude Code built, give `small`'s review Step its own [`harness` and `model`](/docs/workflows/#harnesses-and-models), and the rest of the Workflow stays on Claude Code:

```toml
# In a copy of .vago/workflows/small.toml
[[steps]]
id      = "review"
name    = "Review"
type    = "agent"
prompt  = "prompts/small-review.md"
harness = "codex"
model   = "gpt-6-sol"
```

A Step that sets `harness` sets `model` too, since a model belongs to one Harness. The same works for the Goal review, with `id = "goal-review"`, or any other agent Step.

The starter prompts work on either Harness. `vago init` puts the `vago-` skills they run in `.agents/skills/`, where Codex reads them, as well as in `.claude/skills/`.

## Caveats

- `codex` must be on your `PATH`, logged in, at a version whose `codex exec` has `--approve-for-me`.
- A Step without its own [`args`](/docs/workflows/#args) gets the Workflow's. If the Workflow sets `args` for Claude Code, give the Codex Step its own, or `args = []`.
- Codex reports no cost. Vago prices a Codex Run from its tokens and the model's list price, and shows no cost for a model it has no price for.