Context Engineering Resources

April 25, 2026 · View on GitHub

A curated collection of resources on context engineering — the practice of designing, structuring, retrieving, and managing the context that large language models and AI agents operate over. Covers tooling (MCP servers, context stores, retrieval layers), agent skills, educational material, and awesome lists.

Last Updated: April 2026

Highlights

A handful of repositories stand out by community traction in this space:

Context7 — up-to-date code documentation injected into LLM context for AI coding tools.
OpenViking — open-source context database unifying agent memory, resources, and skills via a filesystem paradigm.
Agent Skills for Context Engineering — large library of agent skills focused on context management for production agents.
Figma Context MCP — feeds Figma layout data to coding agents like Cursor.
WeKnora — Tencent's RAG framework for deep document understanding and context-aware retrieval.
Context Engineering Intro — popular hands-on introduction centred on Claude Code.

Agent Skills
Awesome Lists
Education
Tools
Related Indices

Agent Skills

Skill packs and plugins designed to be loaded into AI coding agents (Claude Code, Cursor, Codex, etc.) to improve how they handle context.

Agent Skills for Context Engineering

A comprehensive library of agent skills covering context management, multi-agent architectures, and production agent systems.

ContextKit

A Claude Code context-engineering and planning system aimed at individual AI development workflows.

Context Engineering Kit

Hand-crafted Claude Code skills focused on improving agent output quality. Also compatible with OpenCode, Cursor, Antigravity, and Gemini CLI.

Awesome Lists

Curated reading lists and surveys.

Awesome Context Engineering

A comprehensive survey of context engineering — papers, frameworks, and implementation guides spanning prompt engineering through production AI systems.

Awesome LLM Long Context Modeling

Must-read papers and blogs on long-context modeling for LLMs.

Prompt in Context Learning

Resources for in-context learning and prompt engineering, with continually updated material on ChatGPT-class LLMs.

Education

Handbooks, courses, and worked examples for learning context engineering as a discipline.

Context Engineering (Handbook)

A first-principles handbook inspired by Karpathy and 3Blue1Brown — moves beyond prompt engineering into context design, orchestration, and optimization.

Context Engineering Intro

Hands-on introduction to context engineering centred on Claude Code, with patterns transferable to other coding assistants.

Tools

MCP servers, context stores, retrieval layers, memory systems, libraries, and platforms.

.context

A Git-native, AI-optimized documentation system that turns a repo into a living knowledge base for agents.

ACAT (Assistive Context-Aware Toolkit)

Intel's Assistive Context-Aware Toolkit — context-driven assistive communication software (originally built for users with motor neurone disease). Included as an adjacent example of context-aware system design rather than LLM context engineering proper.

Agentic Context Engine

Lets agents learn from experience and persist that learning across runs. Also offered as a hosted product at kayba.ai.

Airweave

Open-source context retrieval layer designed to sit underneath AI agents.

Better Context

Pulls up-to-date context on the libraries and frameworks used in your project so coding agents work from current docs rather than stale training data.

Claude Context

Code-search MCP for Claude Code that makes the entire codebase available as context to any coding agent.

CodeGraphContext

MCP server and CLI that index a local codebase into a graph database, exposing structural code context to AI assistants.

Context

Local-first documentation system for AI agents.

Context-Engine

MCP-based agentic context compression suite.

Context Gateway

An agentic proxy that adds instant history compaction and context-optimization tools to any AI agent workflow.

Context Hub

Andrew Ng's context-hub — a context platform for AI applications.

Context Infrastructure

A context and memory system for AI coding agents — persistent memory, personal rules, skills, and scheduled observations.

Context Mode

Context-window optimisation for AI coding agents — sandboxes tool output for a reported ~98% reduction; works across 12 platforms.

Context Portal (ConPort)

A memory-bank MCP server that builds a project-specific knowledge graph for IDE-based RAG and context-aware development.

Context7

Up-to-date code documentation for LLMs and AI code editors, available as an MCP server and hosted platform.

ContextGem

Library for effortless LLM-driven structured extraction from documents.

ContextKeep

MCP server providing long-term memory storage for AI agents.

Continuous Claude v3

Context-management system for Claude Code — uses hooks, ledgers, and handoffs to maintain state, plus MCP execution and orchestration with isolated context windows per agent.

dotai

A context manager intended to work across multiple agent frameworks.

Figma Context MCP

MCP server that exposes Figma layout information to AI coding agents (Cursor, Claude Code, etc.) so they can act on design files.

FreeTodo

Auto-manages personal task context with AI — a task store designed to be agent-readable.

Headroom

A context-optimization layer for LLM applications — sits between the app and the model to manage what gets sent.

Long Context Attention (USP)

Unified (a.k.a. hybrid 2D) sequence-parallel attention implementation for training and inference of long-context transformer models.

Lossless Claw

LCM (Lossless Context Management) plugin for OpenClaw — preserves agent context across sessions without truncation.

MCP Server for Qdrant

Official Qdrant Model Context Protocol server — exposes a Qdrant vector store as a context source for MCP-aware agents.

OpenContext

Personal context store for AI agents and assistants — capture, search, and reuse project knowledge across Codex, Claude, OpenCode, etc., with a desktop GUI.

OpenViking

Open-source context database designed for AI agents (e.g. OpenClaw). Unifies management of memory, resources, and skills via a filesystem paradigm with hierarchical context delivery.

TrustGraph

Context-development platform — store, enrich, and retrieve structured knowledge with graph-native infrastructure, semantic retrieval, and portable context cores.