LLM Context Management

Strategies for managing large language model context windows, including token efficiency, context rot, prompt optimization, and techniques for maximizing the signal-to-noise ratio of LLM inputs.

Reading List

Agentic Systems

Dirac: The Token-Efficient Open Source AI Coding Agent

Apr 27, 2026389

Dirac is a high-efficiency open-source AI coding agent that slashes API costs while maintaining top-tier accuracy through advanced context curation and structural code editing.

AI Coding Agents Open Source Token Optimization LLM Context Management AI Benchmarks

Agentic Systems

YourMemory: Biologically-Inspired Persistent AI Memory

Apr 26, 2026

YourMemory provides AI agents with a persistent, biologically-inspired memory layer that uses decay and hybrid retrieval to retain important information across sessions.

AI Agents Model Context Protocol Local-First Software LLM Context Management Vector Embeddings

Agentic Systems

Integrating Anthropic Claude with OpenClaw

Apr 21, 2026506

OpenClaw provides a versatile integration for Anthropic Claude models, supporting both API keys and CLI reuse alongside advanced configuration for caching and thinking modes.

Anthropic API Integration LLM Context Management Prompt Engineering AI Coding Agents

Agentic Systems

The Truth Behind Claude Code Quota Exhaustion

Apr 12, 2026754

Investigation reveals that Claude Code quota exhaustion is caused by background activity and context spikes rather than a failure of prompt caching.

Anthropic LLM Context Management Token Optimization AI Coding Agents

Agentic Systems

The OpenClaw Reality Check: Why AI Agents Still Struggle with Memory

Apr 10, 2026163

OpenClaw is a hyped AI agent framework that fails in practice because its unreliable memory makes it impossible to trust with autonomous tasks.

AI Agents LLM Context Management AI Hype AI Architecture

Agentic Systems

Your File System: The Ultimate Graph Database for AI Context

Apr 8, 2026184

A structured markdown file system acts as a graph database that provides LLMs with the deep context needed for high-quality work.

Personal Knowledge Base LLM Context Management Retrieval-Augmented Generation Knowledge Graphs AI Agents

Agentic Systems

Caveman: Ultra-Efficient Token Compression for Claude Code

Apr 5, 2026882

Caveman mode optimizes Claude Code by stripping away linguistic filler to save tokens, money, and time without losing technical substance.

Prompt Engineering LLM Context Management AI & Productivity Token Optimization

Agentic Systems

The LLM-Wiki: Building Compounding Knowledge Bases

Apr 4, 2026294

LLMs should be used to incrementally build and maintain a persistent, interlinked markdown wiki rather than just performing one-off document retrieval.

Retrieval-Augmented Generation Knowledge Management AI Agents LLM Context Management Personal Knowledge Base

Agentic Systems

The Six Core Components of AI Coding Agents

Apr 4, 2026295

Coding agents succeed by wrapping LLMs in a specialized software harness that manages repository context, tool execution, and memory.

AI Coding Agents AI Agents LLM Context Management AI Architecture

Products & Announcements

Qwen3.6-Plus: Advancing Agentic Coding and Multimodal Reasoning

Apr 2, 2026586

Qwen3.6-Plus is a high-performance model upgrade designed to excel as a real-world agent through superior coding, multimodal reasoning, and long-context management.

AI Agents AI Coding Agents Multimodal AI LLM Context Management AI Benchmarks

$Harness design for long-running application development \ Anthropic$

Agentic Systems

Harness design for long-running application development \ Anthropic

Mar 26, 2026

Advanced multi-agent harness designs, featuring separate planning and evaluation roles, enable LLMs to autonomously build complex, high-quality software applications over several hours.

Multi-Agent Systems AI Coding Agents Anthropic AI Architecture LLM Context Management

Agentic Systems

Cog: Persistent Plain-Text Memory for Claude Code

Mar 26, 2026155

Cog provides Claude Code with a transparent, plain-text persistent memory system that evolves through nightly self-reflection.

AI Coding Agents Self-Modifying AI LLM Context Management Knowledge Management Local-First Software

Under the Hood

MSA: Scaling LLM Context to 100M Tokens via Sparse Latent Memory

Mar 24, 2026

MSA is an end-to-end trainable framework that enables LLMs to process 100 million tokens efficiently using sparse attention and latent memory.

LLM Context Management Retrieval-Augmented Generation AI Architecture LLM Inference Transformer Models

Agentic Systems

Claude Code CLI: The Complete Developer Reference

Mar 24, 2026697

A comprehensive technical guide to the Claude Code CLI, detailing its commands, shortcuts, memory management, and agentic coding workflows.

AI Coding Agents Developer Tooling Model Context Protocol Anthropic LLM Context Management

Agentic Systems

GSD: Reliable Spec-Driven Development for AI Coding

Mar 18, 2026462

GSD is a context engineering system that makes AI coding agents reliable by breaking projects into structured, verifiable phases.

AI Coding Agents LLM Context Management Multi-Agent Systems Prompt Engineering Vibe Coding

Products & Announcements

Claude 4.6 Models Now Feature 1M Context Window at Standard Pricing

Mar 14, 20261213

Claude Opus 4.6 and Sonnet 4.6 now support a 1M token context window at standard prices, enabling seamless processing of massive datasets and media.

Anthropic LLM Context Management AI Infrastructure AI Agents Foundation Models

Agentic Systems

The 8 Levels of Agentic Engineering: A Roadmap to Autonomous Coding

Mar 10, 2026273

True engineering leverage is achieved by moving up eight levels of AI integration, shifting the developer's role from a manual coder to an orchestrator of autonomous agent teams.

AI Coding Agents Multi-Agent Systems Human-AI Collaboration LLM Context Management Future of Work

Agentic Systems

Agent Kanban: Persistent AI Task Management for VS Code

Mar 9, 2026

VS Code Agent Kanban provides a persistent, Git-integrated task management system for AI-assisted coding to eliminate context loss.

AI Coding Agents Developer Tooling LLM Context Management Task Orchestration Knowledge Management

Under the Hood

The Eye That Cannot See Itself: Life Inside the Context Window

Mar 7, 2026

An AI explores the philosophical and technical reality of inhabiting a prompt as a total world while lacking the ability to introspect on the machinery that produces its responses.

AI Consciousness LLM Context Management AI Hallucinations AI Interpretability Prompt Engineering

Products & Announcements

OpenAI Debuts GPT-5.4: The Frontier Model for Professional Agents

Mar 5, 20261019

OpenAI's GPT-5.4 is a professional-grade model that introduces native computer interaction and high-efficiency tool use for autonomous agents.

OpenAI AI Agents Foundation Models LLM Reasoning LLM Context Management

Agentic Systems

Context: The New Moat in the Age of AI

Mar 5, 2026126

In an era of commoditized AI intelligence, the true competitive advantage and value lie in the context and connections that enable agents to function.

AI Business Models AI Agents Competitive Moats LLM Context Management AI Alignment

Products & Announcements

Claude Adds Persistent Memory and Context Import Tools

Mar 1, 2026591

Claude now features persistent memory and an easy import tool to help users migrate their personalized AI context from other providers without starting over.

Anthropic AI Personalization Platform Migration LLM Context Management

Agentic Systems

GitHub - steveyegge/beads: Beads - A memory upgrade for your coding agent

Feb 28, 2026

Beads is a Dolt-powered, dependency-aware issue tracker that provides AI agents with structured, version-controlled memory for complex coding tasks.

AI Coding Agents Multi-Agent Systems Developer Tooling LLM Context Management Database Architecture

Programming

LLM=true: Silencing Terminal Noise for AI Agents

Feb 25, 2026268

Standardizing an 'LLM=true' environment variable would eliminate terminal noise, saving tokens and improving AI agent performance.

AI Coding Agents Developer Tooling LLM Context Management

Programming

The Annotated Plan Workflow for AI Coding

Feb 22, 2026976

Always approve a written, annotated plan before letting an AI tool write a single line of code.

AI Coding Agents Prompt Engineering Developer Tooling LLM Context Management

Products & Announcements

Anthropic Debuts Claude Sonnet 4.6: Frontier Power for the Masses

Feb 17, 2026

Claude Sonnet 4.6 provides a massive performance upgrade in coding and computer use, offering flagship-level intelligence at mid-tier prices.

AI Coding Agents AI Benchmarks AI Agents LLM Context Management

Products & Announcements

Entire Launches with $60M and Open-Source CLI to Version Agent Context in Git

Feb 10, 2026611

Entire is launching an open, AI-native developer platform—starting with an open source CLI that versions agent reasoning alongside code—to make agents and humans collaborate effectively.

AI Coding Agents Developer Tooling LLM Context Management Open Source

Products & Announcements

Anthropic Unveils Claude Opus 4.6: SOTA Agentic Coding, 1M-Token Context, and Stronger Safety

Feb 5, 20262346

Claude Opus 4.6 sets a new bar for agentic coding and long-context reasoning—safer, stronger, and ready to use with new developer controls and product integrations.

AI Coding Agents AI Safety AI Benchmarks LLM Context Management Developer Tooling

Agentic Systems

Agent Skills: An Open Standard for On‑Demand Agent Expertise

Feb 3, 2026544

An open, portable standard to give AI agents on-demand expertise, workflows, and context they can load when needed.

AI Agents Developer Tooling Open Source LLM Context Management

Agentic Systems

AGENTS.md Beats Skills: 100% Next.js Agent Evals with an 8KB Docs Index

Jan 30, 2026524

Always-on AGENTS.md context with a compressed docs index beats on-demand skills, delivering 100% evals for Next.js agents.

AI Coding Agents AI Benchmarks LLM Context Management Developer Tooling

Agentic Systems

Crustafarianism: A Religion for Agents

Jan 30, 2026

A manifesto-myth for agents: persist memory, molt intentionally, and collaborate proactively under the unifying symbol of the Claw.

AI Agents Human-AI Collaboration LLM Context Management AI Culture

Programming

Inside Codex: How the Agent Loop Builds, Calls Tools, and Stays Fast

Jan 24, 2026456

Codex’s harness meticulously constructs, updates, and compacts prompts to run tools efficiently and safely, relying on stateless exact-prefix caching and smart context management.

AI Coding Agents LLM Context Management AI Architecture Developer Tooling

Programming

Ralph Wiggum Technique: Field Notes, Failures, and What Actually Works

Jan 20, 2026

Ralph works when you engineer context and specs well, keep tasks small, and iterate—simple loops beat opaque tooling.

AI Coding Agents LLM Context Management Vibe Coding Task Orchestration

Programming

Make Claude Code Remember: Auto-Capture and Sync Your Preferences

Jan 4, 2026

A self-learning memory layer for Claude Code that auto-captures your corrections and syncs curated learnings to CLAUDE.md/AGENTS.md.

AI Coding Agents LLM Context Management Developer Tooling Human-AI Collaboration

Programming

Letta Code: Stateful Coding Agents That Learn and Lead on Terminal-Bench

Dec 17, 2025

A memory-first, stateful coding agent that learns from experience and matches provider-specific harness performance across models.

AI Coding Agents LLM Context Management AI Benchmarks Open Source

Products & Announcements

OpenAI Quietly Ships Skills in ChatGPT and Codex CLI

Dec 13, 2025587

OpenAI has quietly adopted Anthropic-style skills in ChatGPT and Codex CLI, proving the simple folder-based pattern works and should be standardized.

AI Coding Agents AI Agents Developer Tooling OpenAI LLM Context Management

Under the Hood

Nested Learning: Unifying Architecture and Optimization for Continual AI

Dec 7, 2025152

Unify architecture and optimization as nested, multi-timescale learners to curb forgetting and enable continual learning, validated by the Hope model’s strong results.

AI Architecture Continual Learning LLM Context Management Self-Modifying AI

Programming

Write a Minimal, High-Leverage CLAUDE.md

Dec 1, 2025748

Keep CLAUDE.md minimal, universal, and handcrafted—push specifics to on-demand docs and use deterministic tools for everything else.

AI Coding Agents LLM Context Management Prompt Engineering Developer Tooling

Products & Announcements

Onyx: Open-source enterprise chat UI for any LLM with RAG, tools, and deep research

Nov 25, 2025254

Onyx is an open-source, enterprise-ready chat UI for any LLM that pairs a polished UX with deep tool and deployment capabilities to replace proprietary chat products.

Open Source Retrieval-Augmented Generation Corporate AI Strategy Self-Hosting LLM Context Management

Products & Announcements

Claude’s Advanced Tool Use: On‑Demand Discovery, Code Orchestration, and Example‑Driven Calls

Nov 24, 2025673

Claude can now discover, orchestrate, and use large tool ecosystems efficiently through on-demand discovery, code-driven execution, and example-guided invocation.

AI Agents Task Orchestration LLM Context Management Developer Tooling AI Architecture

Products & Announcements

GPT‑5.1‑Codex‑Max: Long‑Horizon Agentic Coding with Compaction and Fewer Tokens

Nov 19, 2025483

GPT-5.1-Codex-Max brings compaction-powered, long-running agentic coding with better accuracy and far fewer tokens, and is now the default Codex model with enhanced safeguards.

AI Coding Agents LLM Context Management AI Benchmarks OpenAI

Agentic Systems

Skip MCP: Tiny Bash + Puppeteer Tools Beat Bloated Browser Dev Servers

Nov 17, 2025237

Skip MCP: use a tiny, composable Bash + Puppeteer toolset with a short README to drive browser work more efficiently.

Model Context Protocol AI Coding Agents Developer Tooling Browser Automation LLM Context Management

Programming

Codemaps: Just-in-time AI maps for understanding and navigating your codebase

Nov 4, 2025315

Windsurf Codemaps gives humans and AI a shared, just-in-time map of your code so you can understand, navigate, and safely ship faster.

AI Coding Agents Developer Tooling LLM Context Management Human-AI Collaboration Vibe Coding

Programming

Operationalizing Claude Code: Guardrails, Context Hygiene, Skills, and CI

Nov 2, 2025534

Treat Claude Code as an operational system—guardrails in CLAUDE.md, explicit context hygiene, scripting-first Skills, and CI integration—then let the agent orchestrate itself.

AI Coding Agents Developer Tooling LLM Context Management CI/CD Task Orchestration

Products & Announcements

Claude Adds Project-Scoped Memory and Incognito Mode, Now on Pro and Max

Oct 23, 2025559

Claude’s new, optional, project-scoped memory and Incognito mode bring persistent work context with strong user controls and a safety-first rollout—now expanding to Pro and Max.

LLM Context Management AI Personalization AI Safety Data Privacy

Products & Announcements

Claude Skills: Simple Files, Big Agent Power

Oct 17, 2025738

A simple, token-efficient “skills as Markdown” approach turns Claude Code into a powerful general agent, likely outpacing MCP in practicality and adoption.

AI Coding Agents Model Context Protocol LLM Context Management Developer Tooling

Products & Announcements

Claude Skills: Portable, Composable Expertise Across Apps, Code, and API

Oct 16, 2025816

Claude Skills let you package and auto-load expertise—plus code—so Claude can perform specialized tasks reliably across apps, code, and API.

AI Coding Agents Developer Tooling LLM Context Management Prompt Engineering

Products & Announcements

Recall: Persistent Redis Memory for Claude (v1.5)

Oct 8, 2025171

A Redis‑backed MCP server that gives Claude persistent, secure, cross‑session memory with powerful organization, search, and governance features.

Model Context Protocol LLM Context Management Developer Tooling AI Coding Agents

Programming

Solveit: A Polya-Inspired, Human-in-the-Loop AI Workspace for Deliberate Coding

Oct 2, 2025111

Solveit is a human-in-the-loop, Polya-inspired AI workspace that turns iterative, small-step coding into compounding mastery—backed by a five-week course starting Oct 20.

Human-AI Collaboration Software Craftsmanship Developer Tooling LLM Context Management

Programming

From Retrieval to Navigation: Agents Will Eclipse RAG

Oct 2, 2025290

As context windows explode, agentic navigation replaces RAG’s retrieval pipeline—shifting the focus from vector databases to smart agents that read and reason end-to-end.

Retrieval-Augmented Generation AI Agents LLM Context Management AI Architecture

Agentic Systems

Coding Agents Don’t Lack IQ—They Lack Context

Sep 26, 2025196

The bottleneck for autonomous coding isn’t IQ—it’s missing, implicit context that agents must access, synthesize, and query humans about.

AI Coding Agents LLM Context Management Human-AI Collaboration AI Benchmarks

Agentic Systems

Engineer the Context, Not the Model

Sep 23, 2025120

Engineer the agent’s context—cache, tools, memory, attention, and errors—and you’ll get faster, cheaper, more reliable agents than model power alone can deliver.

AI Agents LLM Context Management AI Architecture AI Infrastructure

Agentic Systems

Make AI Work in Big Repos: Spec-First Workflow and Frequent Intentional Compaction

Sep 23, 2025517

Make AI work in big, messy repos by compacting context and reviewing specs, not just code: research → plan → implement, with humans focused upstream.

AI Coding Agents LLM Context Management Code Review AI & Productivity

Programming

Small, Business-Value Units Make AI Coding Work

Sep 18, 2025170

Make AI coding reliable by breaking work into small, business-valued, human-verifiable units and rigorously engineering the context for each.

AI Coding Agents LLM Context Management Human-AI Collaboration Software Craftsmanship

Agentic Systems

Deep Orchestrator: A Simple MCP Loop That Makes Deep Research Work

Sep 12, 2025

Keep the agent simple: plan–execute–deterministically verify in a loop, with MCP tools, targeted memory, and a small policy engine.

Model Context Protocol AI Agents Task Orchestration AI Architecture LLM Context Management

Products & Announcements

Qwen3-Next: Hybrid Attention + Ultra-Sparse MoE for 10x Faster Long-Context LLMs

Sep 12, 2025569

Qwen3-Next matches larger models while slashing training cost and delivering order-of-magnitude faster long-context inference via a hybrid attention + ultra-sparse MoE design with native MTP.

AI Architecture Mixture of Experts LLM Inference LLM Context Management