Tuning Deep review

⚡ Deep review looks for problems in one layer of a review. It sends the change through several lenses, one reviewer agent per lens. Each lens checks the change from one angle. A final synthesis step merges the findings, ranks them and drops the noise, so you get a short list instead of a pile. Each finding becomes a thread on its line.

Deep review never runs on its own. Select ⚡ Deep review on a layer in the Review Room when you want it. See Reviewing a PR stack.

The five lenses

Turn the last two on for risky changes.

Pick the lenses and their models

1. Open Settings → Deep Review.
2. Use the toggle next to a lens to turn it on or off.
3. Pick a model for each lens. Inherit uses your Screen PRs triage model.
4. Pick a model for Synthesis, the step that merges and ranks the findings.

Each lens can run on a different model. For example, run Soundness & bugs on a cloud model and the other lenses on a local model through Ollama. Local lenses run one after another, because they share one GPU. Cloud lenses run in parallel. A change reviewed entirely on a local model makes no cloud calls.

Rewrite a lens’s instructions

1. Open Settings → Prompts.
2. Find the lens, or the synthesis step, and select Customize.
3. Edit the instructions and select Save.

You change only the instructions. SimpleEdit still adds the output format and the change itself, so an override can’t break a run. There are no placeholders to fill in. If your file is empty or can’t be read, SimpleEdit uses the default instructions for that run.

To go back, select Reset. When SimpleEdit ships new default instructions, a customized prompt is marked Outdated. Select Compare to see the difference.

Give it your team’s rubric

Put a REVIEW.md in the repository: at its root, or in a folder the change touches. Deep review, the overview and the Companion read it as a rubric. Write what your team cares about in a review, for example “Every public function needs a test.” SimpleEdit never creates or edits this file.