Skill UI

This skill is designed to assist developers in debugging complex code issues. It analyzes detailed error messages, stack traces, and common error types (like TypeError or ReferenceError) to accurately pinpoint the root cause of bugs. It provides a structured, step-by-step report including specific fix suggestions and proactive prevention strategies.

A comprehensive guide for designing and implementing robust Infrastructure as Code (IaC) using Terraform or OpenTofu. Covers module hierarchy design, best practices for naming conventions, structuring multi-environment pipelines, advanced testing strategies (native tests, Terratest, static analysis), and integrating security scanning (trivy, checkov) into CI/CD workflows.

This meta-skill runs three essential quality gates (ruff, mypy, and pytest) concurrently against the staged changes. It provides a comprehensive, CI-like check locally, offering a single BLOCK or APPROVE verdict to ensure code readiness before committing. It combines static analysis, type checking, and unit testing into one workflow.

This meta-skill implements a combinator pattern for deep security analysis. It runs three independent gates—policy governance, secret scanning, and audit logging—over a candidate operation. It enforces a strict arbitration rule (Policy DENY > Scanner WARN > ALLOW) to provide a robust, multi-layered verdict, making it ideal for reviewing critical code changes or environment modifications for safety compliance.

This meta-skill is designed to automate the diagnosis of complex programming errors. When presented with a stack trace, traceback, or failing log, it orchestrates a multi-step investigation. It parses the trace, searches the repository history, checks relevant issues, and analyzes code diffs to provide a comprehensive brief for root cause analysis and debugging.

An internal, language-agnostic probe designed for investigating complex stack traces when the source language is unknown. It analyzes the provided trace to output structured, language-neutral elements, including failure contract checks, minimal reproduction guidance, and specific patch targets. This tool is vital for robust failure analysis and debugging across diverse codebases.

A static analysis tool that audits SQL code for common cost and performance anti-patterns (e.g., full-table scans, Cartesian joins). It calculates a query health score (0-100) for data warehouses like BigQuery, Snowflake, and Redshift, providing a prioritized, actionable plan with concrete fixes and estimated cost savings. Essential for FinOps and data quality assurance.

This skill guides developers on implementing Conversational Language Understanding (CLU) using the official azure-ai-language-conversations Python SDK. It demonstrates how to analyze conversation intent and entities, build advanced NLP features, and integrate language understanding into applications using ConversationAnalysisClient. Best practices emphasize using DefaultAzureCredential and context managers for robust, production-ready code.

This advanced skill performs a full-sweep analysis, providing a unified assessment of code quality across multiple dimensions: code decay, architecture integrity, technical debt, and test coverage. It automatically applies safe fixes directly to the codebase, while requiring confirmation for risky changes. It is ideal for comprehensive pre-release code health audits and automated refactoring.

A systematic, non-mutating investigation designed for complex software failures, including bugs, regressions, crashes, and flaky behavior. This skill launches four specialized, read-only sub-agents in parallel to perform deep root-cause analysis, trace code paths, pinpoint the smallest failing seam, and predict the quickest path to confirming or fixing the issue. It is designed purely for diagnosis and hypothesis generation.

A comprehensive, four-phase methodology for debugging complex software issues. It guides users through root cause investigation, pattern analysis, hypothesis testing, and controlled implementation. The core principle emphasizes finding the underlying cause rather than applying quick, symptomatic fixes, ensuring robust and reliable code improvements across various failure scenarios.

A powerful plugin that allows Claude to delegate tasks to various external AI Command Line Interfaces (CLIs), such as OpenAI Codex and Google Antigravity. Use this skill to seek a 'second opinion,' cross-check analysis, or execute side tasks (like research or validation) from multiple model families (Gemini, GPT-OSS) without leaving the main session. It synthesizes and critiques the results, providing a comprehensive, multi-perspective view.