AI Agent News: Releases, Launches and Usage Jumps, Live · page 4
-
Novita Sandbox hits 132.1K weekly downloads, 351% above its average
Novita Sandbox, an SDK for code execution and cloud environments from novita.ai, logged 132.1K downloads in the week ending Nov 15, 2026. That is a new weekly record, well above its recent run rate.
-
Hebbia Chat requests failed for 1 h 15 min in a critical incident
Hebbia, an AI platform that analyses large document sets for finance and legal teams, had a critical incident on Nov 13 in which Chat requests were failing. According to the official status page, it lasted 1 h 15 min.
-
How the Claude Code team thinks about verification
Claude Code team discusses their approach to verification.
-
Running coding agents without handing them my credentials
Article about running coding agents securely without exposing credentials to the agents themselves.
-
GPT-6.2 Astra and Sol now available on AI Gateway
OpenAI's GPT-6.2 Astra and Sol models are now available via Vercel's AI Gateway, with improved document generation and computer use capabilities.
-
Agent downloads up 25% in a week
387K npm downloads in the 7 days to Nov 17, against 309K the week before.
-
Augment Code downloads up 27% in a week
54.3K npm downloads in the 7 days to Nov 17, against 42.6K the week before.
-
Botpress downloads down 31% in a week
24.8K npm downloads in the 7 days to Nov 17, against 36.1K the week before.
-
Claude-Cage gained 4,784 GitHub stars in a week
Now at 6,464 stars.
-
Hermes Agent downloads up 35% in a week
88.2K PyPI downloads in the 7 days to Nov 17, against 65.4K the week before.
-
Novita Sandbox downloads up 49% in a week
120K npm downloads in the 7 days to Nov 17, against 80.9K the week before.
-
Openmotion gained 1,882 GitHub stars in a week
Now at 11.3K stars, against about 106 a week over the month before.
-
Day Trader Skill gained 1,701 GitHub stars in a week
Now at 8,453 stars.
-
dsh-Go gained 7,447 GitHub stars in a week
Now at 26.4K stars.
-
OpenCode 0.6.0 released
opencode-pty 0.6.0 (protocol 9) from packages/process-daemon, published to npm as @opencode/pty@0.6.0.
-
Build memory-aware AI agents with Agentic Retrieve on Amazon Bedrock Managed Knowledge Base
Amazon Bedrock Managed Knowledge Base adds memory capabilities to store conversations and extract facts, enabling multi-user agents with private per-user memory and document access.
-
Oh My Pi 18.22.0 released
Fixed slow streaming of Gemma tool calls with large arguments.; Moved binary analysis from the standalone ida and ghidra tools into the backend-neutral dsm namespace (xd://eval/dsm).; Replaced the ida.enabled, ida.maxOpen, and ida.idleCloseSec settings with…
-
Taxonomy of multi-agent design patterns, anti-patterns, and lessons from manufacturing
AWS published a taxonomy of multi-agent design patterns and anti-patterns with a framework for deciding when multiple agents are needed and how to build cost-efficient systems.
-
GitHub Copilot weekly releases — November 9
GitHub Copilot adds support for GitHub Enterprise Server repositories in the Copilot app and improves agent session visibility in VS Code.
-
OpenCode Browse
-
GitHub Copilot models deprecated on November 16, 2026
GitHub deprecated certain models across Copilot Chat, inline edits, ask and agent modes, and code completions.
-
Control where GitHub Copilot customizations come from
Enterprise administrators can now require GitHub Copilot customizations to come from plugins or managed settings instead of standalone configurations.
-
VS Code retires the Local harness in February 2027
VS Code 1.146 deprecates the Local harness, the older software coordinating model, context, and tools for GitHub Copilot agent sessions in the editor, with retirement planned for February 2027.
-
Fin: Fin and Intercom are down
Critical impact; resolved after 33 min.
-
Amp: Amp is unavailable
Major impact; resolved after 7 min.
-
How MoEngage built agentic customer engagement on Amazon Bedrock AgentCore
MoEngage built Merlin AI, a set of agents on Amazon Bedrock AgentCore that enable marketing teams to manage campaigns through conversation, using MCP tools to access platform APIs with data isolation.
-
Langfuse 4.81.0 released
feat(datasets): filter dataset items by status; feat(ai-gateway): add Google Vertex AI as a gateway provider; feat(trace): show topics transcript on trace via three dots (internal)
-
Introducing explicit prompt caching with open-weight models on Amazon Bedrock
Amazon Bedrock now supports prompt caching for open-weight models, reducing input costs and latency by reusing long context across requests. The post demonstrates caching with the Chat Completions API and an OpenCode coding agent.
-
Show HN: Coral – Open-source gateway where agents write their own tools
Coral is an open-source gateway that enables agents to write their own tools.
-
Retell AI: Call service disruption
Major impact; resolved after 13 min.
-
AI made integrations the vendor’s problem
Report discusses how MCP adoption across vendors has shifted integration responsibilities from users to vendors in AI agent development tools.
-
Claude Agent SDK: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2
Major impact; resolved after 1 h 14 min.
-
Claude Code: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2
Major impact; resolved after 1 h 14 min.
-
Claude Cowork: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2
Major impact; resolved after 1 h 14 min.
-
Claude Design: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2
Major impact; resolved after 1 h 14 min.
-
Claude in Chrome: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2
Major impact; resolved after 1 h 14 min.
-
AI Gateway now supports Private Inference
AI Gateway now supports Private Inference, allowing fine-tuned and self-hosted models to run through unified authentication, SDKs, and observability.
-
Advancing computer-use agents for healthcare teams
OpenAI discusses advances in computer-use agents for healthcare teams.
-
Agents are users, so your software needs evals
Vercel argues that agents are now software users, and developers building agents or agent-compatible software should use evals to measure agent performance across discovery, selection, and access.
-
Chatwoot Captain passes 40K 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.