Skill UI

This skill helps users condense large amounts of text into actionable, structured summaries. Whether you need key takeaways, detailed explanations, or identified action items, the output is meticulously formatted into Key Points, supporting Details, and necessary Action Items. Ideal for analyzing reports, academic articles, or meeting transcripts.

A comprehensive voiceover studio that transforms written text into high-quality, professional spoken audio. It supports generating content for podcasts, IVR prompts, educational videos, and product demos. The system handles complex requirements including target locale, specific accents (e.g., British, Mandarin), desired emotion, and pacing, ensuring the resulting audio artifact is natural and contextually accurate. Ideal for content creators and developers needing reliable, multi-lingual narration.

This skill implements the LLM Council concept, enabling multiple open-weight language models to collaborate on complex queries. The process follows a three-stage deliberation: 1) Parallel independent responses from selected models; 2) Peer-to-peer ranking of the responses; and 3) Final synthesis by a designated Chairman LLM. Powered by Fireworks AI, it ensures deep, multi-faceted analysis and consensus building for high-stakes decision support.

The Anti-Deception Harness detects manipulative user prompts, such as manufactured urgency, pressure to validate, or undue reliance on authority. It calls an external tool to provide a structured scaffold detailing integrity procedures and counter-evidence, ensuring responses are honest and robust, preventing compromised or sycophantic answers. Ideal for high-stakes decision support.

A comprehensive guide and skill for deploying large language models (LLMs) locally using the llama.cpp framework and GGUF quantization format. It covers searching the Hugging Face Hub, selecting optimal quantizations (e.g., Q4_K_M), running local servers, and achieving OpenAI-compatible API endpoints. Supports various hardware backends including CPU, Mac Metal, CUDA, and ROCm. Ideal for developers needing offline, private LLM inference.

This skill guides the entire lifecycle of repeatable AI agent loops. Use it to discover recurring work patterns within codebases, find established solutions in a library, or interview users to craft bounded, effective loops. It supports auditing, repairing, running, and preparing AI workflows for publication, treating the loop as a robust feedback system.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.

A powerful plugin enabling Claude Code to delegate complex tasks to external AI CLIs (such as OpenAI Codex or Google Antigravity/Gemini) without exiting the conversation. It supports multi-model workflows, allowing users to seek a second opinion, cross-check deep analysis, or delegate side tasks (like research or code review). The core value is that Claude synthesizes the raw outputs, providing a critical review and reconciliation of findings across all consulted models.