# Agentic Review

> Refines the code in a change before a human reads it. Runs five facet reviewers per repository, fixes what they find inside the change's own scope, and repeats for up to three rounds.

- Skill: `asermax/agentic-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add asermax/agentic-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/asermax/agentic-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: asermax (https://skillmd.com/u/asermax)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/asermax/agentic-review

---


Load mahou:basics first. Then read `.mahou/agentic-review.md` if present.

# Agentic review

Cleans up a change so mahou:human-review spends its attention on judgment instead of on mechanical findings. It reviews for reuse, simplification, efficiency, altitude, and conformance to the conventions the wiki holds.

It does not look for correctness bugs. Those belong to the human pass, where someone can weigh whether the behavior was intended.

## Fixing without asking

`basics` says every decision belongs to the user and every choice goes back to them as a question. This skill is the exception, and it holds inside these bounds:

- It edits only files the changeset already touches, and creates no new ones.
- It fixes nothing that would change what the code is meant to do.
- It leaves everything uncommitted, so the user can read the whole pass with one `git diff` and discard it as a whole.
- It reports every fix and every skip.

A finding whose fix would break any of those becomes a skipped finding in the report, for the user to rule on.

## Step 1: Establish the changeset

Load mahou:changeset for the repositories, the base, and the diff. Follow the wiki indexes to the entries about the technologies the diff touches; the conventions facet reviews against them and every fix has to follow them, together with the project's `CLAUDE.md` and `.mahou/basics.md` when present.

## Step 2: Dispatch the facet reviewers

Five per repository, one per facet, all in a single message so they run together.

| Facet | Agent | Looks for |
|---|---|---|
| Reuse | `mahou:reuse-reviewer` | new code re-implementing something the codebase already has |
| Simplification | `mahou:simplification-reviewer` | redundant or derivable state, copy-paste with variation, deep nesting, dead code left behind |
| Efficiency | `mahou:efficiency-reviewer` | redundant computation, repeated I/O, independent work run sequentially, closures pinning large scopes |
| Altitude | `mahou:altitude-reviewer` | special cases layered on shared infrastructure where the mechanism underneath should generalize instead |
| Conventions | `mahou:conventions-reviewer` | departures from the wiki entries and the project's standing rules the change touches |

Give each one the repository's absolute path, the base, the diff, and the path to write its findings to. Give the conventions reviewer the path of the project's `CLAUDE.md` as well; it loads mahou:wiki itself.

Reviewers write findings to that file and return the path and a count. Nothing comes back in the message body, which truncates long reports and costs a round trip to recover.

Spawn them unnamed and keep the `agentId` from each spawn result. That is what resumes them in later rounds.

## Step 3: Apply what came back

Read every findings file. Collapse findings that different facets raised about the same line or mechanism, then fix each survivor.

Skip a finding when its fix would change intended behavior, reach outside the bounds above, or when you judge it a false positive. Record the skip and its reason. Never argue with it.

Run the repository's linters and tests afterwards, scoped to the files you touched rather than the whole tree. The project's `CLAUDE.md` usually names the commands; when it does not, detect them from the lockfile or manifest and say which you ran.

## Step 4: Round again, up to three times

Resume each reviewer by its `agentId` and tell it what you fixed. It already holds what it reviewed last round and works out for itself what to re-read.

Stop when a round returns no findings, or after the third round, whichever comes first.

## Step 5: Report and hand off

Per repository: what each round fixed, what it skipped and why, and the linter and test results. Everything stays uncommitted.

Then hand off to mahou:human-review.