wrxck-claude-plugins

Scaffold hardened MCP servers wrapping CLI tools

Create detailed Mermaid diagrams for architecture, flows, ERDs, and more — and improve READMEs with visual documentation

Enforce TypeScript quality - no any types, prefer named exports, no console.log/debugger

Enforce React quality - no inline styles, accessibility checks, semantic HTML

Enforce Jest best practices - proper mock typing, jest.mocked() patterns

C/C++ memory safety and security checks - smart pointers, safe string functions

Code style enforcement - comments, imports, file length, icons/emojis, TODOs, library choices

Security guards - prevent secret leaks and block sudo commands

Database safety - prevent accidental data destruction, migration safety checks

Docker quality and safety - Dockerfile best practices, compose validation, dangerous command blocking

Infrastructure security - nginx, port exposure, infra changes, env sync, hardening audits

Git workflow - commit validation, conventional commits, PR creation, releases, code review

Deploy safety - smoke tests, perf regression detection, test gates, deploy workflow, CI scaffolding

Package management - enforce exact versions, check outdated/vulnerable packages

Learns from your corrections and writes hooks so Claude never repeats the same mistake

Ensures open source packages are properly released on npm, GitHub, and documented

Autonomous security auditor. Scans a GitHub repo for vulnerabilities, triages false positives, writes a PoC, fixes each confirmed bug in its own PR, independently reviews the fix, and merges when the review is clean.

A persistent knowledge brain for Claude Code, backed by Trilium Notes. Standards, decisions, lessons learned, and per-app context that survive across sessions and grow over time. Includes a setup wizard, an MCP server with brain_recall / brain_remember / brain_scan_transcripts tools, and a discipline skill that makes Claude actually use it.

Mines your Claude Code history, config, and session transcripts, spots the workflows you keep repeating, and forges them into the right artifact — a skill, subagent, hook, slash command, workflow, or a full publishable plugin. Local-only analysis; hands off to mcp-forge, behavior-hooks, and matts-second-brain where they fit.