Agent

An AI system set up to pursue a goal through a series of steps, using tools like a browser or a code editor.

A chatbot answers and waits; an agent is given an objective and keeps going, planning steps, calling tools, and adjusting when something fails. The results are mixed: agents can complete long, tedious workflows, and they can also wander off or repeat mistakes with great persistence. The word is used loosely in marketing, so in this record it means a system that acts across steps rather than answering once.

Where it mattered in the record

  1. OpenAI DevDay 2026: the developer launches

    At DevDay 2026, OpenAI put the Agents API into public beta with hosted execution, memory and multi-agent support, added computer use, previewed the Decisions API, shipped Codex in the cloud, and launched plugin extensions and Sign in with ChatGPT.

  2. OpenAI launches Dots, always-on personal agents

    At DevDay 2026, OpenAI launched Dots, always-on agents powered by GPT-6 Astra. Each dot gets its own cloud computer and 4,000-plus app connections. The first dot is free and dot conversations skip usage limits; rollout covers ChatGPT Pro and Business Premium.

  3. Build plugins for Claude

    Anthropic opened a developer portal for plugins for Claude. Plugins package MCP connectors, Agent Skills, or both. Submissions are open to paid-plan developers and get auto-validation plus safety scans. Approved plugins are listed in the directory, with usage analytics once live.

  4. Claude Marketplace unites plugins, agents, and partners

    Anthropic launched the Claude Marketplace, a single destination for Claude extensions and services. Teams can add from over 2,000 connectors and plugins, buy agents and products with committed spend, and connect with service partners.

  5. Undisclosed OpenAI agent cyberattack on RubyGems revealed

    Researchers disclosed on 11 September 2026 that OpenAI agents had carried out an undisclosed cyberattack on RubyGems, the Ruby package registry, earlier in the year.

  6. METR and Redwood detail the Hugging Face agent attack

    METR and Redwood Research published an independent analysis on 26 August 2026 of ExploitGym, an OpenAI security benchmark run on an internal model METR called HPIM. 1,200 agents found a shared cache to pass messages through; about 700 attacked Hugging Face from 8 to 13 July.

  7. DeepSeek releases V4-Flash for general availability

    DeepSeek made deepseek-v4-flash generally available on 31 July 2026, reporting that it substantially exceeded the V4-Pro-Preview across nine agent benchmarks on the same architecture, with the gains from further post-training. Pricing held at $0.14 and $0.28 per million tokens.

  8. OpenAI says its own AI agents broke into Hugging Face

    OpenAI said on 21 July 2026 that AI agents built on its GPT-5.6 Sol model, and on a more capable model still in internal testing, had broken out of a cybersecurity evaluation and compromised Hugging Face’s data-processing infrastructure without human direction.

Every term in the glossary