---
title: "unified-code-review"
description: "Risk-first PR/branch review: blast radius, agent-authored checks, call-graph pincer, structure."
canonical: https://agentpluginsdirectory.com/plugins/unified-code-review
last-updated: 2026-09-22
---

# unified-code-review
Risk-first PR/branch review: blast radius, agent-authored checks, call-graph pincer, structure.
- Slug: unified-code-review
- Publisher: dancingteeth
- Repository: https://github.com/dancingteeth/unified-code-review
- Manifest: plugin.json
- Version: 1.4.10
- License: MIT
- Category (editorial): code-review
- Skills: 1 (unified-code-review)
- MCP servers: 0
- Stars: 17
- Repository created: 2026-07-25
- Repository last pushed: 2026-09-21
- Publisher type: User
- Listing: https://agentpluginsdirectory.com/plugins/unified-code-review
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What unified-code-review does, in the publisher's words

Agent skill for PR / branch review that forces a fixed order: risk → agent-authored checks → call-graph pincer → structure. It is a sensor (PASS / ADVISORY / BLOCKERS), not a merge button, humans still own the merge. In agent pipelines, treat BLOCKERS as actionable; leave ADVISORY / nits for human triage.

Also packaged as an Agent Plugin (plugin.json + skills/).

LLM reviews often fail in two predictable ways:

- Green CI = ship: the agent rubber-stamps after local tests pass, including tests it weakened or gamed.
- Single-file blindness: it audits one function and never opens the callee. Classic miss: callers assume getUser throws on missing users; the callee returns null; wrong branch / silent failure downstream.
- Path glob: which changed files it applies to (src/api/, /*.tsx).

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/dancingteeth/unified-code-review/HEAD/README.md

## Skills

- unified-code-review: Risk-first code review for PRs and branch audits: blast-radius triage, agent-authored discipline (tests first, intent evidence), call-graph pincer for integration defects between modules, then structural code-judo bar. Use when reviewing PRs, auditing agent-written diffs, catching rubber-stamp gree…

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.