Skip to content
Agents tracked: 501 Releases (7d): 777 Agent status: no reported issuesSince last limit reset: Claude 10d · Codex 8d For AI agents: Markdown Updated Nov 21, 2026

AI Agent News: Releases, Launches and Usage Jumps, Live · page 21

  1. Incident Status page

    Claude in Chrome: Elevated errors for Claude Sonnet 5.5

    Major impact; resolved after 42 min.

  2. Release GitHub Releases

    OpenCode 2.2.0 released

    Create, install and share Desktop plugins (experimental); Refreshed the desktop app experience; Add experimental image and decision generation APIs to plugins and sdk, allowing image generation or decision model (think typesafe, clef, etc) workflows.

  3. Release GitHub Releases

    Pydantic AI 2.62.0 released

    Change thinking effort mid-conversation via OpenAI configuration_update and Anthropic per-message output_config, recorded as ModelSettingsDeltaPart; Add ModelRequestRejectedError, ModelAuthenticationError and ModelNotFoundError model error categories; State…

  4. Official Vercel

    Vercel Agent sessions now cost 67% less

    Vercel reduced Agent session costs by 67% through improved context management and better prompt-cache reuse, with updated CLI skills for workflows.

  5. Official Vercel

    HarnessAgent now supports image inputs for all harness adapters

    Vercel's AI SDK HarnessAgent layer now supports image inputs across all harness adapters, allowing users to pass images with text in standard message formats.

  6. Data AgentGid data

    mini-SWE-agent downloads down 35% in a week

    3.1M PyPI downloads in the 7 days to Oct 27, against 4.7M the week before.

  7. Data AgentGid data

    OpenHands downloads down 48% in a week

    53.6K PyPI downloads in the 7 days to Oct 27, against 103K the week before.

  8. Data AgentGid data

    AgentScope downloads up 29% in a week

    61.2K PyPI downloads in the 7 days to Oct 27, against 47.3K the week before.

  9. Data AgentGid data

    Skill downloads up 21447% in a week

    25.2K npm downloads in the 7 days to Oct 27, against 117 the week before.

  10. Data AgentGid data

    Ego Lite passes 20K GitHub stars

  11. Data AgentGid data

    Anatomy gained 3,770 GitHub stars in a week

    Now at 21.6K stars.

  12. Data AgentGid data

    AutoSwarm-Shadow-Adapter gained 1,846 GitHub stars in a week

    Now at 4,000 stars, against about 510 a week over the month before.

  13. Data AgentGid data

    Openbot gained 1,958 GitHub stars in a week

    Now at 2,900 stars, against about 130 a week over the month before.

  14. Data AgentGid data

    Solidcraft gained 2,940 GitHub stars in a week

    Now at 11.3K stars.

  15. AnalysisHot Simon Willison

    GPT-6.1 Terra

    OpenAI announced GPT-6.1 Terra, a model with visual reasoning and computer use capabilities alongside improved writing.

  16. Release GitHub Releases

    Headroom 0.42.0 released

    pipeline: put per-request saved tokens on OutcomeSnapshot (8cedf7b); proxy: let turn hooks rewrite OpenAI Responses requests on HTTP and WebSocket (fbf7bc3); proxy: report the same tool savings on gateway /v1/compress as the proxy path (a904210)

  17. Official GitHub Changelog

    GitHub Copilot assisted permissions generally available

    GitHub Copilot's assisted permissions feature is now generally available across Copilot CLI, app, and VS Code, allowing users to grant or deny permissions when Copilot requests to run commands or access file system paths.

  18. Release GitHub Releases

    eve 0.84.0 released

    4a4951e: Tools and connections now share the Approval policy contract; approvalKey and ApprovalContext.approvedTools are removed.; 4a4951e: Session commands now return Delivery receipts with independent stream() and result() reads.; 4a4951e: Public APIs now…

  19. Official OpenAI

    Greptile takes bug detection beyond code review with GPT-6 Astra

    Greptile uses GPT-6 Astra for testing user workflows, Daybreak for security, and Codex to reduce rework in bug detection.

  20. Hacker NewsHot the-decoder.com

    OpenAI's agents talked to each other through the CPU Cache

    Research describes how OpenAI agents communicated with each other through CPU cache.

  21. IncidentHot Status page

    Synthflow: Platform

    Critical impact; resolved after 1 h 11 min.

  22. Release GitHub Releases

    Oh My Pi 18.13.0 released

    Added SummaryOptions.remotePromptCacheKey and remoteCompactionPolicyModel() so remote compaction can carry a cache key gated against the request it actually sends; Remote compaction now sends cache and session headers for the endpoint and model the request…

  23. OfficialHot GitHub Changelog

    GPT-6.1 Terra in GitHub Copilot

    OpenAI's GPT-6.1 Terra model is now generally available in GitHub Copilot for coding and terminal work at lower cost than GPT-6.1 Sol.

  24. OfficialHot OpenAI

    Introducing GPT-6.1 Terra

    OpenAI released GPT-6.1 Terra, a fast model for computer use, visual reasoning, and writing tasks.

  25. Incident Status page

    Gorgias AI Agent: Ticket assignment notification delivered with delay

    Major impact; resolved after 1 h 7 min.

  26. Official AWS Machine Learning

    Groundhog Day for AI agents: Prompt caching patterns for using OpenAI models on Amazon Bedrock

    Amazon Bedrock now supports prompt caching for OpenAI models, reducing input costs and latency by caching agent context, tools, and instructions across steps.

  27. Incident Status page

    Bits AI SRE: Elevated Error Rates for Workflow Automation, App Builder, Bits Code, Bits Agent Builder

    Major impact; resolved after 56 min.

  28. Official AWS Machine Learning

    Build agentic retrieval on a Managed Amazon Bedrock Knowledge Base with CDK

    AWS CDK template demonstrates building an agent for retrieval across multiple Amazon Bedrock Knowledge Bases with routing, grounding, and observability.

  29. Release GitHub Releases

    E2B 2.61.0 released

    bea0fe6: - Added an idempotencyKey (JavaScript) and idempotency_key (Python) option (+5 more releases that day)

  30. Hacker News github.com

    Show HN: Hob - Open-source, on device alternative to Claude Cowork and ChatGPT

    Hob is an open-source on-device alternative to Claude Cowork and ChatGPT.

  31. Release GitHub Releases

    Junie 3922.1 released

  32. Official OpenAI

    GPT-6.1 Sol helps CodeRabbit strengthen security across complex codebases

    CodeRabbit uses GPT-6.1 Sol to investigate security risks in codebases, with a security agent helping developers review findings and fixes.

  33. Release GitHub Releases

    Codex Security 0.7.0 released

    GitHub Action for Codex Security 0.7.0.; Installs @openai/codex-security@0.7.0.; yaml uses: openai/codex-security@c1551aa2983456a05c616804f311b2d4636280c4 # action-v0.7.0 (+1 more release that day)

  34. Official Vercel

    Vercel Sandbox commands now support stdin

    Vercel Sandbox now supports stdin for running commands, enabling communication with stdio MCP servers like agent-browser.

  35. Limit reset @thsottiaux on X

    OpenAI Codex: usage limits reset

    Reset usage limits today, as announced in a 'Day 21' post.

  36. Official Vercel

    GPT-6.1 Terra now available on AI Gateway

    GPT-6.1 Terra is now available through Vercel's AI Gateway with improvements in visual reasoning and coding for document and desktop application tasks.

  37. Data AgentGid data

    OpenHands downloads down 68% in a week

    39.9K PyPI downloads in the 7 days to Oct 26, against 125K the week before.

  38. Data AgentGid data

    AgentScope downloads up 33% in a week

    60.8K PyPI downloads in the 7 days to Oct 26, against 45.8K the week before.

  39. Data AgentGid data

    OpenShell downloads up 32% in a week

    133K PyPI downloads in the 7 days to Oct 26, against 101K the week before.

  40. Data AgentGid data

    OpenShell passes 20K GitHub stars

This list shows notable items; the sections above show everything of their kind, including patch releases of smaller projects and low-impact incidents. Summaries of outside articles are written automatically from the source's own text and may miss nuance, so follow the link before relying on one.