skills.rest
pear-studio
pear-studioCOMMUNITY
Explore all skills by this author →
GitHub

review1-raise

Generate structured adversarial review findings from a local Git diff.

34|12|Updated Aug 1, 2021
Search Tags:#code review#multi-agent#change analysis#git diff#documentation audit#review workflow#adversarial review
One-click install
npx skills add https://github.com/pear-studio/nonebot-dicepp --skill review1-raise
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review1-raise
Source: https://github.com/pear-studio/nonebot-dicepp/tree/main/docs/agent/skills/review1-raise
Command: npx skills add https://github.com/pear-studio/nonebot-dicepp --skill review1-raise

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python3, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill helps you convert a local Git diff into a clear, structured set of review findings so the next reviewer stages can reply, confirm, and execute with consistent scope and numbered issues.

Core Features & Use Cases

  • Diff-driven review intake: Collects git diff HEAD (with optional user-supplied extra scope) and produces numbered issues (R1, R2, ...).
  • Adversarial, evidence-based critique: Enforces rigorous severity labeling and justification to reduce weak or vague review output.
  • Design-document aware flow: Detects when the target is a design document (e.g., under openspec/changes/) and switches to a design-specific review approach.
  • Multi-agent reference generation: Determines review “stage depth” (light/normal/deep) based on diff size and runs the appropriate number of sub-reviewers, then consolidates into an independent final report.
  • History de-duplication: Reads prior .temp/review-*.md records and avoids repeating issues that were already accepted, stored, or rejected.

Quick Start

Run the review1-raise stage to analyze your local diff and generate a review record for the subsequent steps.

Frequently Asked Questions about review1-raise

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate structured code review findings from a git diff?▼

To generate structured code review findings from a git diff, you can run a multi-agent review workflow that analyzes local diffs and outputs severity-based numbered issues. This process collects history-aware de-duplicated records for subsequent review stages.

What is adversarial code review and how does multi-agent review work?▼

Adversarial code review enforces rigorous severity labeling and evidence-based critique to reduce weak output. Multi-agent review determines stage depth based on diff size, runs appropriate sub-reviewers, and consolidates results into an independent final report.

Can I scope a code review to specific file paths or commit ranges?▼

Yes, you can scope a code review by providing extra user-supplied paths or ranges. The review workflow collects the base git diff and applies your optional scope parameters to focus the adversarial analysis on specific changes.

Does the review workflow support design document changes?▼

Yes, the review workflow supports design document changes by detecting targets like openspec changes. It automatically switches from standard code analysis to a design-specific review approach tailored for document modifications.

How do I avoid duplicate code review issues across multiple review stages?▼

To avoid duplicate code review issues, the workflow reads prior temporary review records and performs history-aware de-duplication. This prevents repeating findings that were already accepted, stored, or rejected in earlier stages.

Do I need Python to run the multi-agent code review workflow?▼

Yes, you need Python3 installed to run the multi-agent code review workflow. The skill relies on Python scripts like the review record tool to orchestrate sub-reviewers and generate the final structured review report.

Related Skills

Explore More Software Engineering Skills→
remix-runremix-run
33.3k

review-pr

Review Remix pull requests from a local checkout for correctness and security.

Official
Intermediate
pear-studiopear-studio
34

review4-execute

Executes finalized consensus implementation entries from a review document after validation and tests.

Community
Advanced
axel-kaliffaxel-kaliff

rethink

Analyze git history and code to identify design flaws and scope drift.

Community
Advanced
loriensleafsloriensleafs

decision-critic

Critique decisions through decomposed claims, evidence verification, and adversarial review.

Community
Advanced