What is the 3-pass auditable review mode?

The 3-pass mode is a systematic process for modules where the AI reads every file three times: first for correctness, second for security, and third for architecture, ensuring complete coverage without skipping files.

How does the confidence scoring filter results?

To ensure accuracy, the skill assigns a confidence score (0-100) to every finding. Only issues with a confidence of 75% or higher are included in the final report to prevent false positives.

Can I save the code review results to a file?

Yes, by including the --file flag in your command, the skill will automatically generate a REVIEW.md file in your project root containing the full report.

What dimensions are covered in the review rubric?

The review evaluates code across six key areas: Code Quality, Security (including OWASP-style vulnerabilities), Architecture, Performance, Testing Gaps, and Documentation.

Lore Code Review

Name: Lore Code Review
Author: LayerDynamics

byLayerDynamics

0•

安全与测试

Performs structured, multi-scope code reviews using a rigorous 3-pass auditable workflow to identify security vulnerabilities and architectural flaws.

Lore Code Review is a specialized skill for Claude Code that enables deterministic and repeatable code audits. It moves beyond simple suggestions by offering four distinct review modes: targeted file review, a comprehensive 3-pass module audit that ensures 100% file coverage, full-project structural analysis, and agent-dispatched topic investigations. Every review follows a standardized rubric covering code quality, security, architecture, performance, testing, and documentation, ensuring that findings are objective, evidence-based, and prioritized by confidence and severity.

主要功能

01Systematic 3-pass auditable workflow for guaranteed file coverage

02Multi-mode review scopes (Files, Module, Project, and Topic-based)

030 GitHub stars

04Comprehensive 6-dimension evaluation rubric including Security and Architecture

05Evidence-based confidence scoring to eliminate AI hallucinations and noise

06Standardized markdown reporting with actionable verification checklists

使用场景

01Conducting targeted reviews of specific file changes before pull request submission

02Generating high-level architectural assessments of entire project codebases

03Performing deep security and logic audits on mission-critical modules

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add layerdynamics/lore review

For use in Claude.ai and ChatGPT