{"slug":"ai-agents","topic":"ai-agents","title":"LLM Agents & Tool Use: State of the Field (October 2026)","summary":"Covers LLM agents that plan and call tools: function calling, coding and computer-use agents, multi-agent systems, benchmarks, security and runtime safeguards, as of October 2026. The key takeaway is that headline agent scores are fragile: benchmark leakage and weak tests can change what a reported number means, and interface drift can erode reliability after deployment, so evaluation design matters as much as model choice.","version":1,"published_at":"2026-10-30","updated_at":"2026-10-30","stats":{"words":3336,"citations":49,"claims":28,"timeline":20,"benchmarks":10,"open_problems":8,"key_papers":15,"glossary":17},"verification":{"status":"verified","method":"independent AI fact-check against cited sources (evidence pack + web checks); automated citation resolution","claims_checked":28,"issues_fixed":21,"verified_at":"2026-10-30"},"ai_assisted":true,"price":{"usd":"0.20","protocol":"x402","network":"eip155:84532","asset":"USDC"},"url":"https://space-kitty.com/dossiers/ai-agents","preview_url":"https://space-kitty.com/api/v1/dossiers/ai-agents/preview","purchase_url":"https://space-kitty.com/api/v1/dossiers/ai-agents","toc":[{"level":2,"text":"Executive summary","anchor":"executive-summary"},{"level":2,"text":"Background and core concepts","anchor":"background-and-core-concepts"},{"level":2,"text":"Timeline of key milestones","anchor":"timeline-of-key-milestones"},{"level":2,"text":"State of the art","anchor":"state-of-the-art"},{"level":3,"text":"Tool calling and API use","anchor":"tool-calling-and-api-use"},{"level":3,"text":"Coding agents and software engineering","anchor":"coding-agents-and-software-engineering"},{"level":3,"text":"Computer-use and web agents","anchor":"computer-use-and-web-agents"},{"level":3,"text":"Multi-agent systems","anchor":"multi-agent-systems"},{"level":3,"text":"Runtime safeguards and monitoring","anchor":"runtime-safeguards-and-monitoring"},{"level":3,"text":"Quantitative benchmarks","anchor":"quantitative-benchmarks"},{"level":2,"text":"Open problems and active debates","anchor":"open-problems-and-active-debates"},{"level":2,"text":"Major programs and players","anchor":"major-programs-and-players"},{"level":2,"text":"Outlook: what to watch (next 12-18 months)","anchor":"outlook-what-to-watch-next-12-18-months"},{"level":2,"text":"Key papers","anchor":"key-papers"},{"level":2,"text":"Glossary","anchor":"glossary"}],"sample_claims":["ReAct reports absolute success-rate gains of 34 and 10 points over imitation and reinforcement learning baselines, respectively, on ALFWorld and WebShop.","Toolformer learns when and how to call external tools from self-generated supervision."],"includes":["markdown","claims","timeline","benchmarks","open_problems","key_papers","glossary","bibliography"],"citation_count":49,"how_to_buy":"GET https://space-kitty.com/api/v1/dossiers/ai-agents with an x402 client ($0.20, USDC on Base Sepolia (testnet))."}