Skill UI

Cavecrew provides a set of specialized subagents (investigator, builder, reviewer) designed for efficient code development delegation. Unlike vanilla agents, Cavecrew compresses output results, drastically reducing the token usage injected back into the main context. This allows users to maintain context over much longer, complex coding sessions, effectively preventing context exhaustion during large-scale refactoring, debugging, or multi-step feature development.

Use this comprehensive skill to generate, manipulate, and export draw.io diagrams. It converts diagrams into native .drawio files and allows exporting to high-quality formats like PNG, SVG, and PDF. Crucially, it embeds the editable XML data, ensuring the resulting diagrams are professional, scalable, and easily integrated into documents.

A code-generation toolkit used for dependency injection in Go. Wire resolves the dependency graph entirely at compile time, emitting plain Go constructor calls without relying on runtime containers or reflection. This approach significantly improves type safety, performance, and error detection by shifting dependency resolution failures to the build phase. Ideal for building robust, scalable backend services in Go.

This skill guides the development of robust, Unix-like Command Line Interfaces (CLIs) in Go using the industry-standard `cobra` library. It covers the entire lifecycle of a CLI application, including command tree setup, managing persistent vs. local flags, advanced positional argument validation, and utilizing the full execution hook chain (RunE). Ideal for engineers adopting or maintaining sophisticated Go CLI projects.

A powerful utility that converts Markdown (`.md`) files into richly formatted Word (`.docx`) documents. Built using pure JavaScript, it requires no external dependencies like Pandoc or LibreOffice. Features include automatic generation of title pages (from YAML front-matter), dynamic table of contents, and accurate embedding of PNG images, making it ideal for technical documentation and reports.

This skill analyzes the comprehensive commit history between two branches to generate a concise, objective, and highly structured Pull Request (PR) description. It automatically groups changes by domain, summarizing the purpose, technical notes, and user impact, ensuring consistency and minimizing manual documentation effort.

This skill assesses the AI readiness of a codebase by running a comprehensive audit against multiple technical and governance pillars. It generates a self-contained, detailed HTML dashboard that provides an overall maturity score, identifies strengths, and, most importantly, delivers a prioritized remediation plan (Fix First / Fix Next). Ideal for auditing repositories before major AI feature development.

This skill generates comprehensive, tailored instruction files for AI coding assistants like GitHub Copilot and Claude. It supports multiple strategies (flat or nested) and is specifically designed for managing complex monorepos by creating area-scoped instructions using `applyTo` globs. Use it to standardize AI tooling usage across large codebases.

This skill analyzes a user's recent activity history captured by the local screenpipe daemon. It detects repeated research or operational workflows, clusters these patterns, and compares them against existing skills. It then drafts new skills or suggests composition recipes to cover unaddressed workflows, helping the user formalize their operational processes.

This utility evaluates complex arithmetic expressions using arbitrary-precision decimal mathematics. It is specifically designed for scientific, financial, or engineering calculations where standard floating-point arithmetic causes unacceptable rounding errors (e.g., 0.1 + 0.2). Supports addition, subtraction, multiplication, division, exponents, and parentheses.

A deep-dive session designed to stress-test a technical plan against an existing domain model. This process forces the user to resolve ambiguities, sharpen terminology, and validate decisions by cross-referencing code and existing documentation (CONTEXT.md, ADRs). It ensures architectural rigor and completeness before implementation.

A persistent knowledge base that transcends standard RAG by integrating three core layers: raw source storage, a structured knowledge graph, and a dynamic wiki layer. It ingests documents once, automatically extracting entities and relations to build a comprehensive, cross-referenced, and verifiable knowledge graph. Ideal for enterprise research, complex Q&A, and deep knowledge accumulation.