Skill UI

This skill provides expertise in designing the critical interface between Large Language Models (LLMs) and external tools. It focuses on tool schema design, function calling best practices, and robust error handling to prevent common failures like hallucination or silent loops. It ensures that the tool's documentation and schema are crystal clear, allowing the LLM to use complex external APIs reliably and predictably.

This comprehensive guide serves as a blueprint for building reliable, production-grade AI applications powered by LLMs. It moves beyond simple demos, covering critical real-world challenges such as prompt optimization for cost reduction, structured output validation, handling hallucinations, and building robust safety and defense layers. Learn to treat prompts as code and ensure enterprise-level reliability.

Expert guidance for fine-tuning LLMs with Axolotl, covering YAML configs, LoRA/QLoRA/DPO variants, multimodal support, and practical troubleshooting patterns for development and training workflows.

Expert guidance for fine-tuning LLMs with Axolotl, covering YAML configs, LoRA/QLoRA/DPO variants, multimodal support, and practical troubleshooting patterns for development and training workflows.

Bridges frozen vision encoders and LLMs for high-quality image captioning, VQA, retrieval, and multimodal chat, enabling zero-shot reasoning without fine-tuning and offering efficient Q-Former training.

Behavioral rules for LLMs writing or reviewing code, emphasizing caution, simplicity, surgical edits, and goal-driven execution to avoid common mistakes during implementation tasks, bug fixes, or refactors.

A comprehensive framework for building persistent memory systems within Large Language Models (LLMs). It supports multi-faceted memory management, including short-term context storage, long-term knowledge retrieval, and entity-based fact updates. This skill is crucial for developing sophisticated AI assistants that maintain coherence, recall user history over extended periods, and provide context-aware interactions by managing memory consolidation and retrieval relevance.

Evaluates large language models across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag) with standardized prompts and metrics, making it easy to benchmark HuggingFace or API models, compare releases, and track training progress.

Fine-tune LLMs with TRL’s RLHF toolkit, covering SFT, DPO, PPO/GRPO and reward model training so you can align HuggingFace Transformers with human preferences or feedback in post-training workflows.

Integrates Apple's FoundationModels into iOS apps so developers can run on-device LLMs for text generation, guided structured outputs via @Generable, custom tool calls, and snapshot streaming without leaving the device.

Fuzzing Dictionary supplies domain-specific tokens to guide fuzzers when targeting parsers, protocols, configs, or file formats so they reach deeper logic; include magic bytes, headers, keywords, and combine generation from LLMs, headers, or binaries before passing via the fuzzer flag.

This guide details the use of the GGUF format and quantization techniques for running large language models (LLMs) efficiently. GGUF enables optimal inference on consumer hardware, Apple Silicon, and systems requiring CPU-only operation. By compressing models using various K-quant methods (like Q4_K_M), developers can significantly reduce memory footprint and hardware requirements while maintaining high performance.