[{"data":1,"prerenderedAt":431},["ShallowReactive",2],{"blog-tag-en-context-engineering":3},[4,83,131,213,313,381],{"slug":5,"published":6,"minutes":7,"category":8,"tags":9,"tagPages":15,"keywords":17,"about":29,"sources":39,"cover":73,"og":74,"expertise":75,"locales":76,"lang":77,"title":80,"description":81,"coverAlt":82},"context-rot-compaction-long-running-agents","2026-10-16",10,"agents",[10,11,12,13,14],"Context rot","Compaction","Context engineering","Long-running agents","Agent memory",[16],"context-engineering",[18,19,20,21,22,23,24,25,26,27,28],"context rot","context rot LLM","compaction AI agents","context engineering","long-running agents context window","Claude compaction API","context editing clear tool results","OpenAI responses compaction","lost in the middle LLM","agent context management","sub-agents context isolation",[30,33,36],{"name":31,"url":32},"Context window","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FLarge_language_model#Context_window",{"name":34,"url":35},"Attention (machine learning)","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FAttention_(machine_learning)",{"name":37,"url":38},"Intelligent agent","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FIntelligent_agent",[40,43,46,49,52,55,58,61,64,67,70],{"title":41,"url":42},"Chroma: Context Rot: How Increasing Input Tokens Impacts LLM Performance (Jul 2025)","https:\u002F\u002Fresearch.trychroma.com\u002Fcontext-rot",{"title":44,"url":45},"Liu et al.: Lost in the Middle: How Language Models Use Long Contexts (2023)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.03172",{"title":47,"url":48},"Modarressi et al.: NoLiMa: Long-Context Evaluation Beyond Literal Matching (2025)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.05167",{"title":50,"url":51},"Anthropic: Effective context engineering for AI agents (Sep 2025)","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Feffective-context-engineering-for-ai-agents",{"title":53,"url":54},"Anthropic: Effective harnesses for long-running agents (Nov 2025)","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Feffective-harnesses-for-long-running-agents",{"title":56,"url":57},"Claude API docs: Compaction at a token threshold","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fbuild-with-claude\u002Fcompaction-threshold",{"title":59,"url":60},"Claude API docs: Compaction on demand","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fbuild-with-claude\u002Fcompaction-on-demand",{"title":62,"url":63},"Claude API docs: Context editing","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fbuild-with-claude\u002Fcontext-editing",{"title":65,"url":66},"Claude API docs: Memory tool","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fagents-and-tools\u002Ftool-use\u002Fmemory-tool",{"title":68,"url":69},"OpenAI API docs: Compaction","https:\u002F\u002Fplatform.openai.com\u002Fdocs\u002Fguides\u002Fcompaction",{"title":71,"url":72},"Cognition: Don't Build Multi-Agents","https:\u002F\u002Fcognition.ai\u002Fblog\u002Fdont-build-multi-agents","\u002Fimages\u002Fblog\u002Fcontext-rot-compaction-long-running-agents\u002Fcover.webp","\u002Fimages\u002Fblog\u002Fcontext-rot-compaction-long-running-agents\u002Fog.jpg","ai-engineer",[77,78,79],"en","de","hu","Context rot in long-running agents: clearing, compaction, memory and sub-agents","Why LLM accuracy drops as context grows, and how clearing tool results, compaction, notes and sub-agents keep long-running agents on track.","Context rot in long-running agents: a robot sits under a pile of papers while a press squeezes a stack into one sheet.",{"slug":84,"published":85,"minutes":86,"category":87,"tags":88,"tagPages":92,"keywords":93,"about":102,"sources":106,"cover":125,"og":126,"expertise":75,"locales":127,"lang":77,"title":128,"description":129,"coverAlt":130},"zep","2026-10-06",11,"rag",[14,89,90,12,91],"Knowledge graph","Temporal graph","RAG",[16,87],[94,95,96,97,98,99,100,101],"zep ai","zep agent memory","graphiti knowledge graph","zep pricing","zep vs mem0","long-term memory for agents","temporal knowledge graph","zep cloud",[103],{"name":104,"url":105},"Zep","https:\u002F\u002Fwww.getzep.com\u002F",[107,110,113,116,119,122],{"title":108,"url":109},"Zep pricing: plans, credits and limits","https:\u002F\u002Fwww.getzep.com\u002Fpricing",{"title":111,"url":112},"Zep documentation","https:\u002F\u002Fhelp.getzep.com\u002F",{"title":114,"url":115},"Graphiti on GitHub","https:\u002F\u002Fgithub.com\u002Fgetzep\u002Fgraphiti",{"title":117,"url":118},"Graphiti product page","https:\u002F\u002Fwww.getzep.com\u002Fplatform\u002Fgraphiti\u002F",{"title":120,"url":121},"Announcing a new direction for Zep's open-source strategy","https:\u002F\u002Fwww.getzep.com\u002Fblog\u002Fannouncing-a-new-direction-for-zeps-open-source-strategy\u002F",{"title":123,"url":124},"Graphiti: temporal knowledge graphs for AI agents (arXiv)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.13956","\u002Fimages\u002Fblog\u002Fzep\u002Fcover.webp","\u002Fimages\u002Fblog\u002Fzep\u002Fog.jpg",[77,78,79],"Zep review: agent memory on a temporal graph","Zep: a hosted agent-memory API on a temporal knowledge graph. Credits on writes, free retrieval, Flex from $125\u002Fmonth, Graphiti as the self-hostable part.","Zep review: a robot studies a wall chart of linked cards, each card with a tiny clock.",{"slug":132,"published":133,"updated":134,"minutes":135,"category":8,"tags":136,"tagPages":140,"keywords":142,"about":153,"sources":161,"cover":207,"og":208,"expertise":75,"locales":209,"lang":77,"title":210,"description":211,"coverAlt":212},"ai-agent-memory-design","2026-09-10","2026-10-02",13,[137,12,138,139],"AI agent memory","Memory poisoning","GDPR",[16,141],"gdpr",[143,144,145,146,147,148,149,150,151,152],"AI agent memory design","long-term memory for AI agents","episodic semantic procedural memory LLM","agent memory architecture","ChatGPT memory vs Claude memory","Claude memory tool","AI memory poisoning","LLM context compaction","AI agent memory GDPR","short-term vs long-term memory agents",[154,155,158],{"name":37,"url":38},{"name":156,"url":157},"General Data Protection Regulation","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FGeneral_Data_Protection_Regulation",{"name":159,"url":160},"Prompt injection","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPrompt_injection",[162,164,166,168,171,174,177,180,183,186,189,192,195,198,201,204],{"title":163,"url":66},"Anthropic docs: Memory tool",{"title":165,"url":63},"Anthropic docs: Context editing",{"title":167,"url":51},"Anthropic Engineering: Effective context engineering for AI agents",{"title":169,"url":170},"Sumers et al.: Cognitive Architectures for Language Agents (CoALA)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2309.02427",{"title":172,"url":173},"Packer et al.: MemGPT, Towards LLMs as Operating Systems","https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.08560",{"title":175,"url":176},"Park et al.: Generative Agents, Interactive Simulacra of Human Behavior","https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.03442",{"title":178,"url":179},"Unit 42: When AI Remembers Too Much, persistent behaviors in agents memory","https:\u002F\u002Funit42.paloaltonetworks.com\u002Findirect-prompt-injection-poisons-ai-longterm-memory\u002F",{"title":181,"url":182},"From Untrusted Input to Trusted Memory: A Systematic Study of Memory Poisoning Attacks in LLM Agents (preprint)","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2606.04329v1",{"title":184,"url":185},"The Hacker News: ChatGPT macOS flaw could have enabled long-term spyware via memory function","https:\u002F\u002Fthehackernews.com\u002F2024\u002F09\u002Fchatgpt-macos-flaw-couldve-enabled-long.html",{"title":187,"url":188},"Vectorize: OWASP ASI06, Memory and Context Poisoning explained","https:\u002F\u002Fvectorize.io\u002Farticles\u002Fowasp-asi06",{"title":190,"url":191},"Claude Help Center: Use chat search and memory to build on previous context","https:\u002F\u002Fsupport.claude.com\u002Fen\u002Farticles\u002F11817273-use-claude-s-chat-search-and-memory-to-build-on-previous-context",{"title":193,"url":194},"OpenAI Help Center: Memory in ChatGPT","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F8590148-memory-faq",{"title":196,"url":197},"OpenAI Help Center: Dots privacy, security, and safety FAQs","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F20001529-dots-privacy-security-and-safety-faqs",{"title":199,"url":200},"Flavio Copes: A deep dive into OpenAI dots (quotes the dots documentation on memory)","https:\u002F\u002Fflaviocopes.com\u002Fopenai-dots\u002F",{"title":202,"url":203},"GDPR Article 5: Principles relating to processing of personal data","https:\u002F\u002Fgdpr-info.eu\u002Fart-5-gdpr\u002F",{"title":205,"url":206},"GDPR Article 17: Right to erasure","https:\u002F\u002Fgdpr-info.eu\u002Fart-17-gdpr\u002F","\u002Fimages\u002Fblog\u002Fai-agent-memory-design\u002Fcover.webp","\u002Fimages\u002Fblog\u002Fai-agent-memory-design\u002Fog.jpg",[77,78,79],"Designing memory for AI agents: tiers, write rules, poisoning and GDPR","How to design AI agent memory: context vs session vs long-term tiers, what to write and never store, retrieval, compaction, poisoning and GDPR erasure.","AI agent memory design: a robot files an index card into the open drawer of a three-drawer cabinet.",{"slug":214,"published":215,"updated":216,"minutes":217,"category":8,"tags":218,"tagPages":222,"keywords":224,"about":236,"sources":246,"cover":307,"og":308,"expertise":75,"locales":309,"lang":77,"title":310,"description":311,"coverAlt":312},"token-saving-tools-coding-agents-top-20","2026-08-26","2026-09-23",17,[219,220,12,221],"Claude Code","token usage","developer tools",[223,16],"claude-code",[225,226,227,228,229,230,231,232,233,234,235,21],"reduce Claude Code token usage","token saving tools for coding agents","rtk token killer","lean-ctx","context-mode MCP","Serena MCP","Repomix compress","MCP tool search","prompt caching","ccusage","coding agent cost",[237,240,243],{"name":238,"url":239},"Large language model","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FLarge_language_model",{"name":241,"url":242},"Claude (language model)","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FClaude_(language_model)",{"name":244,"url":245},"Prompt engineering","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FPrompt_engineering",[247,250,253,256,259,262,265,268,271,274,277,280,283,286,289,292,295,298,301,304],{"title":248,"url":249},"rtk: a CLI proxy that filters shell output for coding agents (GitHub)","https:\u002F\u002Fgithub.com\u002Frtk-ai\u002Frtk",{"title":251,"url":252},"lean-ctx: context read modes and shell compression for coding agents (GitHub)","https:\u002F\u002Fgithub.com\u002Fyvgude\u002Flean-ctx",{"title":254,"url":255},"context-mode: sandboxed tool output with SQLite FTS5 search (GitHub)","https:\u002F\u002Fgithub.com\u002Fmksglu\u002Fcontext-mode",{"title":257,"url":258},"Serena: semantic code retrieval and editing over MCP (GitHub)","https:\u002F\u002Fgithub.com\u002Foraios\u002Fserena",{"title":260,"url":261},"token-savior: symbol index, memory and bash compaction over MCP (GitHub)","https:\u002F\u002Fgithub.com\u002FMibayy\u002Ftoken-savior",{"title":263,"url":264},"code-review-graph: a code graph for blast-radius reviews (GitHub)","https:\u002F\u002Fgithub.com\u002Ftirth8205\u002Fcode-review-graph",{"title":266,"url":267},"Aider documentation: repository map","https:\u002F\u002Faider.chat\u002Fdocs\u002Frepomap.html",{"title":269,"url":270},"Repomix: pack a repository into one AI-friendly file (GitHub)","https:\u002F\u002Fgithub.com\u002Fyamadashy\u002Frepomix",{"title":272,"url":273},"Context7: up-to-date library documentation for LLMs (GitHub)","https:\u002F\u002Fgithub.com\u002Fupstash\u002Fcontext7",{"title":275,"url":276},"claude-context: hybrid code search MCP (GitHub)","https:\u002F\u002Fgithub.com\u002Fzilliztech\u002Fclaude-context",{"title":278,"url":279},"caveman: terse output modes for coding agents (GitHub)","https:\u002F\u002Fgithub.com\u002FJuliusBrussee\u002Fcaveman",{"title":281,"url":282},"claude-token-efficient: an eight-rule CLAUDE.md (GitHub)","https:\u002F\u002Fgithub.com\u002Fdrona23\u002Fclaude-token-efficient",{"title":284,"url":285},"claude-code-router: a local model gateway for coding agents (GitHub)","https:\u002F\u002Fgithub.com\u002Fmusistudio\u002Fclaude-code-router",{"title":287,"url":288},"ccusage: token and cost reports from local agent logs (GitHub)","https:\u002F\u002Fgithub.com\u002Fryoppippi\u002Fccusage",{"title":290,"url":291},"LLMLingua: prompt compression (Microsoft, GitHub)","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FLLMLingua",{"title":293,"url":294},"ComputingForGeeks: tools that reduce Claude Code token usage, tested (April 2026)","https:\u002F\u002Fcomputingforgeeks.com\u002Freduce-claude-code-token-usage-tools\u002F",{"title":296,"url":297},"Claude Code docs: manage costs effectively","https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcosts",{"title":299,"url":300},"Claude Code docs: MCP, output limits and tool search","https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fmcp",{"title":302,"url":303},"Claude API docs: tool search tool","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fagents-and-tools\u002Ftool-use\u002Ftool-search-tool",{"title":305,"url":306},"Claude API docs: prompt caching","https:\u002F\u002Fplatform.claude.com\u002Fdocs\u002Fen\u002Fbuild-with-claude\u002Fprompt-caching","\u002Fimages\u002Fblog\u002Ftoken-saving-tools-coding-agents-top-20\u002Fcover.webp","\u002Fimages\u002Fblog\u002Ftoken-saving-tools-coding-agents-top-20\u002Fog.jpg",[77,78,79],"Top 20 ways to cut coding-agent tokens: rtk, lean-ctx, Serena and more, ranked by evidence","rtk, lean-ctx, context-mode, Serena and 16 more token savers for coding agents, ranked by evidence, with my own measurements on a real Nuxt codebase.","Cutting coding-agent tokens: a press shrinks a tall coin stack into a short one, a robot with a trophy beside it.",{"slug":314,"published":315,"updated":316,"minutes":317,"category":8,"tags":318,"tagPages":322,"keywords":324,"about":335,"sources":341,"cover":375,"og":376,"expertise":75,"locales":377,"lang":77,"title":378,"description":379,"coverAlt":380},"multi-agent-systems-when-worth-it","2026-08-07","2026-09-04",12,[319,320,321,12],"Multi-agent systems","AI agents","Orchestrator-worker",[323,16],"ai-agents",[325,326,327,328,329,330,331,332,333,334],"multi-agent systems when worth it","multi-agent vs single agent","orchestrator-worker pattern","multi-agent LLM token cost","why multi-agent systems fail","don't build multi-agents","subagents context isolation","multi-agent debate worth it","agent handoffs vs subagents","when to use multiple AI agents",[336,339,340],{"name":337,"url":338},"Multi-agent system","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FMulti-agent_system",{"name":37,"url":38},{"name":238,"url":239},[342,345,348,351,354,357,360,363,366,369,372],{"title":343,"url":344},"Anthropic Engineering: How we built our multi-agent research system","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fmulti-agent-research-system",{"title":346,"url":347},"Anthropic: Building effective agents","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fbuilding-effective-agents",{"title":349,"url":350},"Cognition: Don't Build Multi-Agents (12 June 2025)","https:\u002F\u002Fcognition.com\u002Fblog\u002Fdont-build-multi-agents",{"title":352,"url":353},"Cognition: Multi-Agents: What's Actually Working (22 April 2026)","https:\u002F\u002Fcognition.com\u002Fblog\u002Fmulti-agents-working",{"title":355,"url":356},"Google Research: Towards a science of scaling agent systems","https:\u002F\u002Fresearch.google\u002Fblog\u002Ftowards-a-science-of-scaling-agent-systems-when-and-why-agent-systems-work\u002F",{"title":358,"url":359},"arXiv 2512.08296: Towards a Science of Scaling Agent Systems","https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.08296",{"title":361,"url":362},"arXiv 2503.13657: Why Do Multi-Agent LLM Systems Fail? (MAST)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.13657",{"title":364,"url":365},"arXiv 2305.14325: Improving Factuality and Reasoning in Language Models through Multiagent Debate","https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.14325",{"title":367,"url":368},"arXiv 2502.08788: Stop Overvaluing Multi-Agent Debate","https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.08788",{"title":370,"url":371},"OpenAI Agents SDK: Handoffs","https:\u002F\u002Fopenai.github.io\u002Fopenai-agents-python\u002Fhandoffs\u002F",{"title":373,"url":374},"Claude Code documentation: Subagents","https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fsub-agents","\u002Fimages\u002Fblog\u002Fmulti-agent-systems-when-worth-it\u002Fcover.webp","\u002Fimages\u002Fblog\u002Fmulti-agent-systems-when-worth-it\u002Fog.jpg",[77,78,79],"Multi-agent systems: when they beat one agent, and when they do not","Orchestrator-worker, fan-out, critic, handoff: what multi-agent systems really buy you, what they cost in tokens, how they fail, and a table to decide.","Multi-agent systems: one robot carries a small box, four robots carry a long beam, a scale between them.",{"slug":382,"published":383,"updated":384,"minutes":385,"category":8,"tags":386,"tagPages":390,"keywords":392,"about":399,"sources":405,"cover":425,"og":426,"expertise":75,"locales":427,"lang":77,"title":428,"description":429,"coverAlt":430},"agents-md-skills-mcp-cli-decision-matrix","2026-06-19","2026-07-16",8,[12,387,388,389],"AGENTS.md","Agent Skills","MCP",[16,391],"mcp",[21,387,393,394,18,395,396,397,398],"agent skills","MCP vs CLI","agent context window","token budget for AI agents","when to use MCP instead of a CLI","sub-agents for AI agents",[400,401,404],{"name":238,"url":239},{"name":402,"url":403},"Model Context Protocol","https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FModel_Context_Protocol",{"name":244,"url":245},[406,407,410,413,416,419,422],{"title":50,"url":51},{"title":408,"url":409},"Chroma: Context Rot (Jul 2025)","https:\u002F\u002Fwww.trychroma.com\u002Fresearch\u002Fcontext-rot",{"title":411,"url":412},"AGENTS.md: the open format for agent instructions","https:\u002F\u002Fagents.md",{"title":414,"url":415},"Agent Skills specification","https:\u002F\u002Fagentskills.io\u002Fspecification",{"title":417,"url":418},"Anthropic: Advanced tool use, tool search and programmatic tool calling (Nov 2025)","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fadvanced-tool-use",{"title":420,"url":421},"Blocks.ai: MCP vs CLI, the context window cost","https:\u002F\u002Fblocks.ai\u002Fblog\u002Fmcp-vs-cli-context-window-cost",{"title":423,"url":424},"Thoughtworks Technology Radar: Context engineering (Adopt, Apr 2026)","https:\u002F\u002Fwww.thoughtworks.com\u002Fradar\u002Ftechniques\u002Fcontext-engineering","\u002Fimages\u002Fblog\u002Fagents-md-skills-mcp-cli-decision-matrix\u002Fcover.webp","\u002Fimages\u002Fblog\u002Fagents-md-skills-mcp-cli-decision-matrix\u002Fog.jpg",[77,78,79],"Context engineering for coding agents: AGENTS.md, skills, MCP or CLI?","Context engineering decides what a coding agent has in context: rules in AGENTS.md, procedures in skills, and when an MCP tool beats a shell command.","AGENTS.md, skills, MCP or CLI: a robot at a crossroads reads a signpost with four arrows: page, toolbox, plug and terminal.",1793451238333]