Skip to content

Lazybone/Codewright

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Codewright

A Claude Code plugin with multi-agent skills for automated code analysis, bug fixing, refactoring, and project audits.

Installation

In Claude Code:

/plugin marketplace add Lazybone/Codewright
/plugin install codewright@Lazybone-Codewright

Skills

audit-project

Runs a comprehensive project audit with 5 parallel subagents (Security, Bugs, Code Hygiene, Structure, GitHub Issues). Each finding is automatically created as a GitHub Issue.

/codewright:audit-project

Requirements: Git repository with GitHub remote, gh CLI installed and authenticated.


codebase-doctor

Analyzes the entire codebase with 7 parallel agents, automatically fixes found issues, and creates a final report. Works in 3 waves: Analysis → Auto-Fix → Review & Verify.

/codewright:codebase-doctor

Modes: report-only, fix-critical, fix-all (default)


github-issue-fixer

Systematically fixes a specific GitHub Issue: load issue, locate bug, create fix plan, implement, verify, and commit.

/codewright:github-issue-fixer

Usage: Pass an issue number or URL as argument.


refactor-orchestrator

Orchestrates a full project refactoring with autonomous subagents. A teamleader agent analyzes the project, creates a plan, and delegates to specialized worker agents.

/codewright:refactor-orchestrator

Configuration: Scope, aggressiveness (conservative/moderate/aggressive), dry-run mode.

codebase-onboarding

Analyzes a codebase and generates architecture documentation and getting-started guides. 3 agents: Structure Scanner → Architecture Analyzer → Doc Writer.

/codewright:codebase-onboarding

Output: ARCHITECTURE.md and/or GETTING-STARTED.md


test-engineer

Finds missing tests, identifies coverage gaps, and writes tests automatically. 2-wave approach: analyze coverage & risk, then generate and review tests.

/codewright:test-engineer

Modes: report-only, critical-paths, full-coverage (default)


pr-reviewer

Multi-perspective code review of a Pull Request using 3 parallel agents (Logic, Security, Quality). Optionally posts review as GitHub PR comment.

/codewright:pr-reviewer

Usage: Pass a PR number or URL as argument.


perf-analyzer

Identifies performance bottlenecks using up to 4 parallel agents (Bundle, Query, Runtime, Infra). Smart agent selection based on project type.

/codewright:perf-analyzer

Output: PERFORMANCE-REPORT.md with impact-sorted findings.


Requirements

  • Claude Code with plugin support
  • Git
  • Optional: GitHub CLI (gh) for audit-project and github-issue-fixer

License

MIT

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages