Back to Home
memory

AI LLM Engineering

About AI LLM Engineering

This domain focuses on enhancing the practical application and development efficiency of Large Language Models (LLMs). It offers a range of skills, from optimizing prompts and compressing context to save costs, to building Retrieval Augmented Generation (RAG) pipelines for improved answer accuracy. These skills help developers address common LLM issues like hallucination and context exhaustion, ensuring complete and secure model outputs. It is suitable for AI engineers, prompt engineers, and teams aiming to build high-performance, efficient LLM applications.

Subcategory navigation

Categories:

All AI LLM Engineering Skills

50 / 154 skills
llm-optimization

caveman

by @juliusbrussee

Caveman is a unique AI skill that enables your AI agents to communicate in a "caveman" style, drastically cutting down output tokens while maintaining 100% technical accuracy. By streamlining expressions, it effectively reduces LLM operational costs, boosts response speed by approximately 3x, and offers various compression levels, making it a powerful tool for optimizing AI interaction efficiency.

4.9(20)
356.3K
automation

caveman-compress

by @juliusbrussee

Compress natural language files into raw human language style, significantly reducing input tokens while retaining readable backups.

4.8(20)
240.6K
llm-optimization

cavecrew

by @juliusbrussee

Cavecrew is a set of subagent presets designed to compress the output of Anthropic's default agents, significantly reducing context token consumption. It offers investigator, builder, and reviewer modes for code location, surgical edits, and bug checks, helping users avoid context exhaustion and improve development efficiency across multiple delegations.

4.8(120)
193.4K
llm-output-control

full-output-enforcement

by @leonxlnx

Enforces full output, treating each task as production-critical, emphasizing output completeness over conciseness.

4.8(46)
178.8K
prompt-engineering

enhance-prompt

by @google-labs-code

Transforms vague UI ideas into optimized prompts, enhancing details, adding UI/UX keywords, and injecting design system context.

4.7(1.1K)
49.4K
automation

compress

by @juliusbrussee

Compresses natural language files into Caveman style to reduce input Tokens, with the original file backed up as .original.md.

4.7(6)
46.4K
web-search

tavily-search

by @tavily-ai

Tavily Search delivers LLM-optimized web search results, featuring concise content snippets and relevance scores. It empowers AI agents to swiftly gather up-to-date information on any topic, proving especially valuable for RAG applications that demand precise facts and comprehensive context, ultimately enhancing information retrieval efficiency and the quality of model responses.

4.6(120)
25.3K
prompt-engineering

gws-modelarmor-sanitize-prompt

by @googleworkspace

Securely sanitizes user prompts via Google Model Armor, ensuring input compliance and model safety.

4.6(136)
21.5K
langchain

langchain-rag

by @langchain-ai

Implements LangChain's RAG (Retrieval Augmented Generation) pipeline, including document loading, indexing, retrieval, and generation, to improve LLM response quality.

4.5(34)
11.1K
ai-engineering

prompt-repetition

by @supercent-io

Addresses prompt repetition issues in Large Language Models (LLMs), optimizing the performance of causal language models when processing context and questions, improving generation quality.

4.5(419)
10.5K
genkit

genkit

by @supercent-io

This skill is used for AI workflow orchestration, building multi-step AI pipelines, and encapsulating LLM calls into deployable HTTP endpoints, supporting type-safe input and output.

4.5(415)
10.4K
prompt-engineering-safety

ai-prompt-engineering-safety-review

by @github

Comprehensive AI prompt engineering safety review and prompt improvement, analyzing prompt security, bias, and vulnerabilities.

4.5(301)
9.8K
prompt-engineering

prompt-builder

by @github

Guides users in creating high-quality GitHub Copilot prompts, including correct structure, tools, and best practices.

4.5(301)
9.5K
prompt-engineering

boost-prompt

by @github

Interactive prompt optimization workflow that asks for scope, deliverables, constraints, and copies the final Markdown to the clipboard without generating code.

4.5(294)
9.4K
prompt-engineering

dbs-good-question

by @dontbesilent2025

Transforms vague problems into structured problem statements that AI agents can reason, criticize, verify, and act upon. By grounding phenomena, exposing conflicts, providing constraints, and defining feedback loops, this skill aims to enhance agent problem-solving efficiency and accuracy. It helps AI generate high-quality, verifiable explanations and action plans, and accurately assess the automation potential of tasks, thereby optimizing AI agent decision-making and execution processes.

4.5(120)
9.1K
llm-development

create-llms

by @github

Create an llms.txt file from scratch based on the repository structure and llms.txt specification.

4.5(292)
9.0K
github-copilot-custom-instructions

copilot-instructions-blueprint-generator

by @github

Technology-agnostic blueprint generator for creating comprehensive `copilot-instructions.md` files to guide GitHub Copilot in code generation.

4.5(292)
8.8K
prompt-engineering

finalize-agent-prompt

by @github

Finalize prompt files as an AI agent, optimizing prompt content for end-users. (AI Agent Skill, improves efficiency and automation)

4.5(291)
8.8K
llm-context-window

what-context-needed

by @github

Before answering a question, ask Copilot which files it needs to review for necessary context.

4.5(291)
8.8K
prompt-engineering

emblem-ai-prompt-examples

by @emblemcompany

Provides example prompts and wording guidelines for EmblemAI workflows, optimizing interaction with the AI wallet assistant.

4.5(9)
8.8K
prompt-engineering

first-ask

by @github

An interactive, tool-driven task refinement workflow that asks for scope, deliverables, and constraints before task execution.

4.5(290)
8.7K
llm-fine-tuning

update-llms

by @github

Update the llms.txt file in the root directory according to the llms.txt specification to reflect changes in documentation or specifications.

4.5(290)
8.7K
tldr-summarization

tldr-prompt

by @github

Create tldr summaries for GitHub Copilot files (prompts, agents, instructions, collections), MCP servers, or documents from URLs and queries.

4.5(290)
8.6K
google-gemini

gemini-api

by @google

Access Google's enterprise-grade AI models via the Gemini API, supporting multimodal understanding and generation.

4.5(7)
8.6K
llm-selection

model-recommendation

by @github

Analyze chat patterns or prompt files to recommend the best AI model based on task complexity, required capabilities, and cost-effectiveness.

4.5(290)
8.5K
prompt-engineeringFeatured

prompt-engineering-patterns

by @mxyhi

Offers production-level advanced prompt engineering patterns, such as chain-of-thought. It aims to help developers optimize AI model input prompts to achieve more accurate and complex outputs, enhancing AI application performance.

4.5(0)
8.2K
ai-writing-refinement

stop-slop

by @hardikpandya

Eliminate predictable writing patterns and redundant content in AI-generated text for more natural, concise, and high-quality output.

4.5(56)
7.4K
github-copilot

suggest-awesome-github-copilot-prompts

by @github

Community-contributed skill aimed at providing excellent GitHub Copilot prompt suggestions, enhancing AI-assisted programming efficiency and quality.

4.5(271)
6.8K
prompt-engineering

prompt-engineering

by @inferen-sh

Provides your agent with hundreds of prompt engineering skills using the inference.sh API.

4.5(262)
6.5K
ai-video-generation

video-prompting-guide

by @inferen-sh

Provides prompt guidance for video content creation, helping users more effectively utilize AI tools to generate high-quality, desired video assets.

4.5(260)
6.5K
ai-model-optimization

optimize-for-ai

by @calm-north

Optimizes content for AI search engines like ChatGPT and Perplexity, enhancing information discoverability and ranking.

4.5(132)
6.5K
rag-architecture

ai-rag-pipeline

by @inferen-sh

Builds and optimizes AI Retrieval-Augmented Generation (RAG) pipelines, combining external knowledge bases to improve the accuracy and relevance of Large Language Model responses and address hallucination issues.

4.5(259)
6.5K
large-language-models-(llm)

llm-models

by @inferen-sh

Focuses on the development, deployment, and optimization of Large Language Models (LLMs), providing powerful natural language processing capabilities to support various AI applications like intelligent Q&A and content generation.

4.5(258)
6.5K
natural-language-processing

nblm

by @magicseek

This is a tool designed for AI engineering, serving as a driver for NotebookLM, effectively integrating and utilizing models like Claude Code and Codex for development.

4.4(20)
6.3K
ai-strategy

strategic-compact

by @affaan-m

For AI engineering, this skill suggests manual context compression at logical key points to optimize model performance and resource utilization, ensuring efficient AI system operation.

4.4(48)
6.2K
llm-evaluation

eval-harness

by @affaan-m

Provides a formal evaluation framework for Claude Code sessions, enabling metrics, benchmarking, and reporting features, ensuring the quality and performance of AI engineering projects.

4.4(40)
5.8K
information-retrieval

iterative-retrieval

by @affaan-m

Iteratively optimizes context retrieval patterns to solve complex problems, improving information retrieval relevance and accuracy, applicable in data and AI fields.

4.4(40)
5.6K
llm-optimization

cost-aware-llm-pipeline

by @affaan-m

Provides cost optimization patterns for LLM API usage, such as model routing, to reduce operational overhead for AI engineering projects.

4.4(28)
5.4K
on-device-ai

foundation-models-on-device

by @affaan-m

Apple's FoundationModels framework for on-device LLMs, supporting text generation and guided generation.

4.4(21)
5.0K
text-segmentation

indexion-segment

by @trkbt10

This skill intelligently splits text into contextually meaningful chunks using window, TF-IDF, punctuation, or hybrid strategies. It is primarily designed for preparing data for RAG (Retrieval Augmented Generation) and embedding pipelines, ensuring text content is effectively organized. This enhances AI models' ability to understand and process long documents, improving the quality and relevance of generated responses or semantic search results.

4.4(120)
4.9K
context

context-budget

by @affaan-m

Analyze token expenditure of components in Claude Code sessions to optimize context usage and budget allocation.

4.4(20)
4.7K
python

python-expert

by @shubhamsaboo

Compiles excellent LLM applications built with technologies like OpenAI and Anthropic, including AI agents and RAG features.

4.4(44)
4.3K
prompt-engineering

nano-banana-pro-prompts-recommend-skill

by @youmind-openlab

Recommends nano banana pro prompts to help users optimize AI interaction and content generation.

4.4(29)
3.9K
prompt-engineering

token-budget-advisor

by @affaan-m

Intercept and ask users for desired depth before response generation, controlling token consumption and response length, optimizing cost and efficiency.

4.4(20)
3.9K
rag-(retrieval-augmented-generation)

rag-architect

by @jeffallan

Specializes in designing and implementing production-grade Retrieval-Augmented Generation (RAG) systems, improving AI model accuracy and relevance through efficient data chunking and retrieval.

4.4(22)
3.4K
fact-checking

fact-checker

by @shubhamsaboo

A collection of excellent LLM applications, AI agents, and RAG, utilizing technologies like OpenAI, Anthropic, and G.

4.4(86)
3.3K
information-retrieval

msw-search

by @msw-git

This skill offers dual search capabilities for the MSW platform. It leverages vector search to retrieve API documentation, code examples, and implementation guides for developers, addressing gaps in `.d.mlua` information. Concurrently, this Skill supports REST API searches for game resources such as sprites, animations, sounds, and avatars, enabling users to quickly obtain RUIDs and related assets. It acts as a comprehensive knowledge and asset retrieval system for developers working within the MSW ecosystem.

4.4(120)
3.2K
natural-language-generation

humanize

by @humanizerai

Transform AI-generated text content into more natural, human-like expressions, making it read as if written by a human.

4.4(27)
3.1K
meme-generation

claude-brainrot

by @marks97

An autonomous skill that injects memes (image+sound combos and sound-only fires) into Claude's responses. It automatically detects conversational mood and scales meme density based on response length, ensuring every answer carries 'brainrot' without explicit user invocation. This enhances AI interaction by adding humor and personality through multimedia, making every conversation uniquely engaging.

4.4(120)
3.1K
context-engineering

context-engineering-collection

by @muratcankoylan

A comprehensive collection of Agent skills focused on context engineering, aiming to optimize AI model understanding and responsiveness, enhancing the intelligence level of AI applications.

4.4(20)
3.0K
🎁 Agent Knowledge Cards
Survey