Run any Skill in Manus with one click

Get Started

configure-reviews

Stars49

Forks3

UpdatedFebruary 23, 2026 at 19:51

Set up DeepWork Reviews — automated code review rules using .deepreview config files

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

Unsupervisedcom

Unsupervisedcom/deepwork

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software DevelopersComputer and Mathematical Occupations·SOC 15-1252

SKILL.md

readonly

name	configure_reviews
description	Set up DeepWork Reviews — automated code review rules using .deepreview config files

Configure DeepWork Reviews

Help the user set up automated code review rules for their project.

How to Use

Read the README_REVIEWS.md file in this plugin directory for the complete reference on DeepWork Reviews
Look at existing .deepreview files in the project that might have something related to what the user wants to do. Reuse existing rules / instructions if possible, including refactoring inline instructions that need to be reused into their own files in the .deepwork/review directory
Ask any clarifications needed using AskUserQuestion. Skip if you are clear on the need.
Create or update .deepreview YAML config files in the appropriate locations in their project
Test your change. Make a small change that should trigger the affected rule, then call mcp__deepwork__get_review_instructions and verify that the output includes text referencing your new rule. Be sure to revert the change.
Summarize your changes.
Ask the user if they'd like you to run a review now using the /review skill to try out the new configuration.

Important

Always read README_REVIEWS.md first (if you have not in this conversation already) — it contains the full configuration reference, examples, and glob pattern guide
Help users think about WHERE to place .deepreview files (project root for broad rules, subdirectories for scoped rules)
Write practical review instructions that will produce actionable feedback
Minimize reviewer count. Each review rule spawns a separate sub-agent with material overhead (context setup, API round-trips, teardown). Combine rules into a single rule whenever the joint instructions would not overload the reviewer's context. For example, running ruff and mypy on Python files should be one python_lint rule — not two separate rules — because the combined instructions are short and the file set is identical. Only split into separate rules when the instructions are long enough or distinct enough that merging them would degrade review quality.

More from this repository

same repository

new-user

Unsupervisedcom/deepwork

Welcome new users to DeepWork — introduce features, set up reviews, and optionally record a first workflow

2026-04-1549

review

Unsupervisedcom/deepwork

Run DeepWork Reviews on the current branch — review changed files using .deepreview rules

2026-04-1449

create-agent

Unsupervisedcom/deepwork

Creates a new LearningAgent with directory structure, core-knowledge.md, and Claude Code agent file. Guides the user through initial configuration.

2026-04-1249

learn

Unsupervisedcom/deepwork

Runs the learning cycle on all LearningAgent sessions with pending transcripts. Identifies issues, investigates root causes, and incorporates learnings into agent definitions.

2026-04-1249

deepreviews

Unsupervisedcom/deepwork

Reference documentation for DeepWork Reviews — automated code review rules using .deepreview configs and DeepSchema-generated rules

2026-04-1049

deepwork

Unsupervisedcom/deepwork

Start or continue DeepWork workflows using MCP tools

2026-04-1049

Configure DeepWork Reviews

Help the user set up automated code review rules for their project.

How to Use

Read the README_REVIEWS.md file in this plugin directory for the complete reference on DeepWork Reviews

Look at existing .deepreview files in the project that might have something related to what the user wants to do. Reuse existing rules / instructions if possible, including refactoring inline instructions that need to be reused into their own files in the .deepwork/review directory

Ask any clarifications needed using AskUserQuestion. Skip if you are clear on the need.

Create or update .deepreview YAML config files in the appropriate locations in their project

Test your change. Make a small change that should trigger the affected rule, then call mcp__deepwork__get_review_instructions and verify that the output includes text referencing your new rule. Be sure to revert the change.

Summarize your changes.

Ask the user if they'd like you to run a review now using the /review skill to try out the new configuration.

Important

Always read README_REVIEWS.md first (if you have not in this conversation already) — it contains the full configuration reference, examples, and glob pattern guide

Help users think about WHERE to place .deepreview files (project root for broad rules, subdirectories for scoped rules)

Write practical review instructions that will produce actionable feedback

Minimize reviewer count. Each review rule spawns a separate sub-agent with material overhead (context setup, API round-trips, teardown). Combine rules into a single rule whenever the joint instructions would not overload the reviewer's context. For example, running ruff and mypy on Python files should be one python_lint rule — not two separate rules — because the combined instructions are short and the file set is identical. Only split into separate rules when the instructions are long enough or distinct enough that merging them would degrade review quality.