Skill UI

Guides execution of implementation plans by dispatching isolated subagents per task, enforcing a spec-first then quality review loop, and preserving overall coordination context for faster, high-quality iterations.

This skill performs a comprehensive full-site SEO audit, crawling up to 500 pages while respecting robots.txt. It analyzes technical health, content quality, performance metrics, and local signals, delegating to specialized subagents. It generates a scored health report (0-100) with prioritized action plans and optional PDF documentation for enterprise use.

Performs a full GEO audit of any website by discovering business type, crawling prioritized pages, and delegating parallel subagents to evaluate AI visibility, platform readiness, technical infrastructure, EEAT content, and schema markup, then outputs a composite GEO score with prioritized actions.

Spawns N parallel AI subagents that compete on the same task using isolated git worktrees. A coordinator dispatches, monitors, evaluates results by metric or LLM judge, then merges the winning branch. Ideal for parallel optimization, content variations, refactors, and research exploration.

The /hub:spawn command launches N subagents simultaneously, each operating in an isolated git worktree to work on the same task. It supports custom templates (like optimizer or refactorer) and manages the entire parallel execution cycle, ensuring diverse strategies and structured results for maximum value exploration.

Analyzes and generates a self-contained, explorable HTML report detailing usage metrics for Claude Code sessions. It tracks token consumption (input/output), subagent usage, skill calls, and cache efficiency across specified time ranges. This tool helps developers visualize resource waste and pinpoint areas for workflow optimization, ensuring cost-effective and efficient AI development cycles.

This skill enables personal execution of implementation plans task-by-task in the current session without spawning subagents per task. It employs Test-Driven Development (TDD) for verification, tracks progress via a ledger, and resolves conflicts autonomously. Suitable for inline execution choices or when subagent tools are unavailable, it reduces context costs while ensuring quality through final review.

This tool generates production-grade, spec-compliant agent definition files (.md) adhering to the Anthropic 2026 schema. It supports both the creation of new custom subagents and the validation of existing agent files against marketplace compliance rules. Use it when building complex agent architectures, reviewing agent quality, or ensuring agents meet high industry standards.

A robust framework for building complex, long-horizon AI agents using LangGraph. It addresses common pitfalls in advanced agent design, such as unbounded state growth (Virtual FS cleanup) and subagent persona leakage (explicit message overriding). Ideal for tasks requiring multi-step planning, task delegation, and self-correction through reflection loops.

This skill executes implementation plans by assigning a dedicated sub-agent to each independent task. It ensures quality through task-specific reviews and a final branch review. It emphasizes continuous execution, autonomous decision-making, and isolated contexts to prevent context pollution, suitable for structured software development workflows.

Cavecrew provides three subagent presets—investigator, builder, and reviewer—that compress tool results into minimal structured output, reducing main-context usage. Ideal for code search, small edits, and diff reviews in long sessions.

This advanced skill manages complex, large-scale development tasks by coordinating multiple, specialized subagents in parallel. It systematically decomposes the goal into isolated missions, manages resource quotas (token usage), and handles error recovery across multiple files and components. Ideal for large feature builds, system refactoring, or complex applications requiring multiple, concurrent roles (Planner, Builder, Tester).