Skill UI

Speeds up LLM inference by combining speculative decoding with Medusa multiple heads and lookahead techniques so chatbots, code assistants, and other real-time workloads can run with 1.5-3.6× lower latency while preserving target-model quality.

This skill optimizes the performance of TwinMind's meeting AI operations. It fine-tunes transcription accuracy and speed using the advanced Ear-3 model, managing audio quality, and implementing caching strategies. Ideal for developers or system administrators needing to integrate, troubleshoot, or optimize high-volume meeting transcription workflows.

Designs and builds production-ready real-time voice agents with OpenAI Realtime, Vapi, Deepgram, ElevenLabs, LiveKit, and WebRTC, optimizing for latency, audio quality, and responsive UX.

Generate or edit high-quality photos, illustrations, and visual assets using OpenRouter models like Gemini 3.1 Flash and Flux 2 Pro; configure the API key once, then run the scripts for prompts, edits, or batch outputs for presentations, marketing, or general creative needs.

Assess research claims and experimental rigor by evaluating methodology, statistical validity, evidence quality, and biases using GRADE and Cochrane ROB frameworks, plus supporting schematics for complex concepts.

Use Nano Banana 2 plus Gemini 3.1 Pro Preview to iteratively craft publication-quality scientific schematics—neural nets, pathways, or system diagrams—matching document-type quality thresholds.

Helps product managers diagnose whether AI workflows rely on indiscriminate context stuffing or deliberate context engineering, then apply bounded domains, retrieval, and Research→Plan→Reset→Implement practices to improve reliability and output quality.

This skill provides comprehensive techniques for optimizing prompts used with any Large Language Model (LLM). It teaches proven methodologies—including Role definition (R), Context setting (S), Constraints (C), Instructions (I), and Templates (T)—to transform vague inputs into precision-engineered instructions. Users can reliably achieve structured JSON output, implement Chain-of-Thought reasoning, reduce hallucinations, and minimize token usage, ensuring consistent and high-quality results across different AI platforms.

A comprehensive guide to building production-grade evaluation systems for Large Language Models (LLMs). This skill covers advanced methodologies, including Direct Scoring and Pairwise Comparison, alongside critical techniques for mitigating systematic biases (e.g., Position Bias, Length Bias). Learn how to select appropriate metrics and structure prompts to ensure reliable, consistent, and objective quality assessment across various AI applications.

This pattern defines a robust, continuous loop structure for autonomous agents. It integrates crucial components like quality gates, evaluation harnesses, and explicit recovery controls. It guides complex workflows by defining decision paths (e.g., PR control, RFC decomposition) and includes strategies to mitigate common failure modes such as infinite loop churn or cost drift, ensuring measurable progress in autonomous systems.

A specialized skill for generating hyper-realistic, humanized images using Google AI Studio (Gemini). It simulates the imperfections of real photography—such as natural lighting, slight camera shake, and depth of field—to create content that looks genuinely captured by a smartphone. Ideal for creating high-quality assets for social media (influencer style) or educational materials.

Assesses AI-targeted content for Experience, Expertise, Authoritativeness, and Trustworthiness plus structure so teams can score and improve citability.