ai-tools

AI Tools

A comprehensive collection of AI-powered development skills, commands, and standards for building high-quality software with AI assistance.

Overview

This repository contains:

• Skills – Reusable AI prompts and workflows for common development tasks
• Commands – Specialized procedures for specific problem domains
• Agentic Coding Standards – A complete framework for maintaining quality and consistency when working with AI-assisted development

Installation

This repo is a Claude Code plugin marketplace. Install everything as a single bundled plugin:

/plugin marketplace add sean-m-cooper/ai_tools
/plugin install ai-tools@ai-tools
/reload-plugins

The first command registers the marketplace; the second installs the ai-tools plugin (all skills and commands); /reload-plugins applies it to the current session. Once installed, skills are invoked with the plugin namespace, e.g. /ai-tools:code-scorecard. Run /plugin update ai-tools to pull the latest version.

Skills

Each skill is a focused, well-structured prompt designed to guide AI agents (like GitHub Copilot CLI or Claude) through a specific development task.

Available Skills

Skill	Purpose
code-map	Generate AI-optimized YAML code maps (.codemap.yaml) that help AI agents quickly understand repository structure and navigate codebases efficiently
code-scorecard	Audit a codebase across nine quality dimensions on a 0–10 scale, using deterministic metrics from the CodeMetrics.AI tool as evidence; supports --verbose and --explain output modes
feature-writer	Write comprehensive feature implementations with proper architecture, testing, and documentation
root-cause	Systematically analyze bugs, failures, and unexpected behavior to identify root causes through structured investigation of code, data, dependencies, and logs
story-writer	Create well-structured user stories and tasks with clear acceptance criteria and technical requirements
story-splitter	Break down complex stories and epics into smaller, manageable tasks
security-audit	Perform an adversarial security audit covering frontend, backend, auth, database, infrastructure, and third-party integrations with threat modeling and structured findings
story-prefinement	Review and score stories for quality and readiness before sprint planning

Each skill directory contains a SKILL.md file with:

• Purpose and use cases
• Detailed workflows
• Examples and best practices
• Integration guidance

code-scorecard Output Flags

The code-scorecard skill accepts optional flags at invocation (e.g. "run the scorecard against eContract.API.slnx --verbose"). They are additive — combine them for the full breakdown.

Flag	Output
(none)	Default. Scorecard table, deterministic evidence summary (which scores came from JSON vs CSV vs qualitative), and the top 3 issues to fix.
--verbose	Adds the projected score lift if the top issues are fixed, the top offenders per metric, and the three-signal breakdown behind each deterministic dimension.
--explain	Adds a self-contained score-derivation section: filter counts, per-signal values with the threshold rows they matched, the score arithmetic, and what each metric (decomposition ratio, max member cyclomatic complexity, maintainability index) actually means.

You can also pass a specific .sln/.slnx path to scope the audit when a repo contains multiple solutions.

Commands

Specialized procedures for solving specific problem domains or handling complex scenarios.

Command	Purpose
root-cause	Systematic approach to identifying and analyzing root causes of defects and system failures

Agentic Coding Standards

The agentic_coding_standards_example/ directory contains a comprehensive framework for maintaining code quality and consistency when working with AI-assisted development.

Key Modules