Skill UI

Prepares Claude with an ultra-granular, line-by-line code-analysis mode that maps invariants, assumptions, and flows before any vulnerability hunting, keeping context stable and reducing hallucinations during audits.

Prepares Claude with an ultra-granular, line-by-line code-analysis mode that maps invariants, assumptions, and flows before any vulnerability hunting, keeping context stable and reducing hallucinations during audits.

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.

This guide outlines three robust architectural patterns (Direct, Standard, Queue-Based) for integrating and enriching data using the Clay SaaS platform. It helps users select the appropriate design—from simple, ad-hoc integrations to complex, enterprise-scale pipelines—by considering data volume, latency, and required automation level for optimal data flow.

This skill guides the application of Domain-Driven Design (DDD) tactical patterns, including defining aggregates, modeling value objects, managing repositories, and utilizing domain events. It helps developers translate complex domain rules into clean, robust, and behavior-rich code structures, ensuring invariants are explicitly enforced.

A comprehensive guide to implementing Exa neural search at various scales. This resource details three distinct architectures—Direct Search, Cached Search, and full RAG Pipeline—allowing developers to choose the optimal blueprint based on traffic volume, required latency, and complexity. Use this when designing scalable Exa integrations, from simple UI search to advanced AI-powered research agents.

A comprehensive guide to selecting and implementing the optimal Firecrawl architecture pattern for various data ingestion needs. Learn to choose between On-Demand, Scheduled Crawl Pipelines, and Real-Time Ingestion Pipelines based on data volume, required latency, and complexity. This is essential for designing scalable, production-grade scraping infrastructure, including RAG and knowledge base applications.

Mamba is a selective state-space model architecture delivering O(n) inference for million-token sequences, offering hardware-aware Mamba-1 and Mamba-2 variants that beat Transformers on speed without KV cache overhead.

This guide details advanced methods for merging multiple specialized LLMs (e.g., combining math, coding, and chat capabilities) without the need for expensive retraining. Learn various techniques such as Linear Merging, SLERP, Task Arithmetic, TIES-Merging, and DARE. Use mergekit to create high-performing, specialized model variants quickly, significantly improving performance and reducing computational costs.

This skill provides comprehensive tools for implementing granular cost controls for OpenRouter API usage. It allows users to set hard spending caps via per-key limits, track real-time credit balances, and implement client-side middleware for budget enforcement. Use cases include preventing unexpected overspending, managing multi-key environments, and optimizing API calls using cost-saving model variants.

This skill provides a comprehensive guide to understanding OpenRouter's API billing structure. It teaches users how to accurately calculate costs based on prompt (input) and completion (output) tokens. Users can utilize provided scripts to estimate costs, check credit balances, and implement cost-saving strategies like using `:floor` or `:free` model variants. Ideal for budgeting, performance comparison, and production cost tracking.