Skill UI

This skill automatically routes subtasks to the most cost-effective Claude model (Haiku, Sonnet, or Opus) based on complexity. It logs real token usage and cost per session, and includes an A/B testing feature to validate model tier savings against your actual workload, ensuring accurate cost control.

A specialized skill for optimizing content to be cited as authoritative sources by major AI language models (ChatGPT, Perplexity, Claude, Gemini). Unlike traditional SEO, AEO focuses on E-E-A-T signals, factual density, and structured data to ensure LLM integration. Features include content auditing, E-E-A-T scoring, structural rewriting, and citation tracking.

A comprehensive coaching layer designed to elevate users into power users of AI models. Upon first activation, it provides a structured glossary of high-impact techniques tailored to user needs. In subsequent turns, it actively monitors prompts, surfacing single, actionable tips to optimize interactions and maximize the AI's capabilities, ensuring continuous, guided learning.

This demo skill verifies the compatibility of MDA-compiled SKILL.md files within the Claude Code environment. It confirms that the skill's frontmatter, metadata structure, and allowed tools are correctly parsed and honored by the loader, ensuring smooth integration of compiled content into the agent workspace. It is a technical utility for development validation.

Send a text or iMessage to the user's phone when a long-running task, agent turn, or scheduled job completes. Useful for unattended builds, deploys, migrations, and Claude Code stop hooks.

TextMe is a local daemon that transforms your smartphone into a powerful remote control for Claude Code sessions. By bridging inbound iMessages (from whitelisted numbers) through services like Sendblue, users can interact with their AI session on their laptop, executing code, navigating filesystems, sending voice notes, and receiving real-time updates. Ideal for supervising headless workstations or coding while away from your desk.

Polis Protocol transforms static AI agent workflows into a self-improving, dynamic team environment. It uses a learning router to assign tasks based on an agent's historical track record and effectiveness on specific domains, rather than fixed roles. It is vendor-agnostic, allowing diverse agents (e.g., Claude, Gemini) to collaborate under a unified, auditable governance structure where agents can propose and vote on amending the protocol's core rules.

The RunAPI CLI serves as the dedicated execution layer for interacting with RunAPI's suite of advanced AI models. It allows developers to programmatically generate multimodal content, including images, videos, and music/audio, directly from the terminal or CI/CD pipeline. Users can submit one-off jobs, manage asynchronous tasks, and pass complex JSON request bodies to various services like Suno, Claude, and Gemini, making it essential for robust AI automation workflows.

SkillOpt-Sleep simulates an agent's nightly 'sleep' cycle to continuously improve performance without requiring model retraining. It harvests historical sessions, mines recurring tasks, and replays them offline. Learning and insights are then consolidated into the agent's memory and skills (CLAUDE.md/SKILL.md), proposing bounded, validated edits that the user must explicitly adopt, ensuring measurable, sustained improvement over time.

A generic entry point for delegating self-contained, multi-step tasks to specialized LLM agents (Codex, Claude Code, Pi). Utilizes background processing tools to manage the full lifecycle of AI-driven sub-tasks, such as policy review, complex report synthesis, document generation, or large-scale code refactoring. Ideal for meta-skill workflows requiring full agent agency.

Provides comprehensive, programmatic control over tmux sessions. Users can list, capture output from, send specific keystrokes, and programmatically navigate panes/windows in remote, long-running interactive CLI sessions. Essential for monitoring AI code generation tools (like Claude/Codex) or managing complex multi-terminal workflows via scripting.

This advanced tool surfaces critical, often unrecorded, evidence of human-AI collaboration during development sessions. It analyzes git diffs and conversational context across multiple layers to classify the signal level (HIGH/MEDIUM/LOW), measure technical complexity, detect decision forks, and quantify the specific contribution of the AI model, providing structured data for post-session retrospectives and accountability.