# cursor Published articles for cursor. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Preventing Destructive Actions with Step-up Authentication DevFeed: [Preventing Destructive Actions with Step-up Authentication]() Original publisher: [Read original article]() Author: Jessica Temporal Published: 2026-09-17T00:00:00Z Content type: tutorial Language: en Sources: [Auth0 Blog]() Topics: [MFA](), [Auth0](), [Security](), [cursor](), [App](), [Next.js](), [passwords]() Tags: [agentic](), [agents](), [auth0](), [cursor](), [don-t](), [mfa](), [next-js](), [prompt](), [security]() ### AI overview This tutorial explains how to protect a destructive "reset all scores" admin action in a Next.js Agentic Arcade app with Auth0. The implementation combines an admin role check with a fresh MFA step-up challenge before the reset request is sent, helping prevent a stale session from triggering the wipe. ### Source excerpt Cursor built a "reset all scores" feature from a single prompt, gated by a fresh MFA step-up. ## Investigating whether LLMs disproportionately use "spine" in GitHub pull request titles DevFeed: [Investigating whether LLMs disproportionately use "spine" in GitHub pull request titles]() Original publisher: [Read original article]() Author: Hillel Wayne Published: 2026-09-16T19:58:30Z Content type: opinion Language: en Sources: [Newsletter feed for Hillel Wayne's Newsletter]() Topics: [Large Language Model](), [Artificial Intelligence](), [GitHub](), [pull-requests](), [anthropic](), [cursor]() Tags: [2026](), [ai](), [claude](), [cursor](), [github](), [llms](), [pull-requests]() ### AI overview The article investigates whether large language models disproportionately use the word "spine" in GitHub pull request titles. Using GitHub Search as a proxy, it finds a much larger increase in "spine" pull requests than overall pull requests, but does not establish that the pattern comes from one specific model; sampled pull requests were coauthored by Claude or Cursor. ### Source excerpt Earlier this year I worked on a couple of AI-generated TLA+ specifications, and one thing I've noticed was that they all used the word "spine" somewhere. Then I saw some non-TLA+ public projects use "spine" too and got curious if it was a new LLMism (LLisM?). If it is, we should see it appear in a discontinuous jump in the number of code projects that use "spine". The proper way to thoroughly explore this would be via the GitHub archive data dump but that'd take forever and I have a full time job now. Instead, I'm going to use GitHub Search. We can't filter code samples by date but can filter pull requests, which seems like an acceptable enough proxy. Here's a graph of the number of public PRs each year with "spine" in the title: I stopped 2026 at 2026-09-01, meaning the first nine months of this year have seen 20 times more "spine" PRs as all of 2025. Now it could be the case that that GitHub grew, like, a lot in the last year. Maybe the 20x increase is due there being 20x as many pull requests. We can account for this by comparing it against all public PRs:1 If "spine" wasn't an LLisM, we would expect only a 1.5xish increase from 2025, not a 20x increase. One other possible confounder could be that LLMs don't like the word spine but PRs Georg submitted 15,000 changes to a chiropractor repo. I will deal with this issue by ignoring it. Can we figure out which models specifically are obsessed with spines? Let's try breaking the change down by month: Looking at LLM timeline, the only widely-used model released in May was Opus 4.8, which came out too late in the month to explain the spike. GPT 5.5 came out in April, which could be the cause. But I'm not convinced by this. For one, the spine count is still 1.5xing each month before April, and I definitely remember seeing spines as early as February. Also, I spot checked and many of the PRs that use "spine" are coauthored by Claude or Cursor. So I think it's not specific to one LLM, but could be convinced otherwise. Othe ## AI SDK harness layer now supports native subscription authentication DevFeed: [AI SDK harness layer now supports native subscription authentication]() Original publisher: [Read original article]() Author: Felix Arntz Published: 2026-09-14T21:28:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [vercel ai sdk](), [Authentication](), [OAuth](), [codex](), [Claude Code]() Tags: [ai](), [authentication](), [claude-code](), [codex](), [coding-agents](), [copilot](), [cursor](), [gateway](), [oauth](), [opencode](), [sdk]() ### AI overview The AI SDK harness layer now supports authenticating supported coding-agent harnesses through their native subscriptions. It uses explicit provider credentials when available, otherwise falls back to a host-native subscription without requiring code or configuration changes. ### Source excerpt The AI SDK harness layer now supports authenticating harnesses through their native subscriptions, where the underlying harness supports them. The harness layer runs different coding agents through the same HarnessAgent interface, so you can switch agents without changing your application code. No code changes or new settings are required. The direct authentication mode uses explicit provider environment credentials when they are present, and otherwise a native subscription found on the host. The default auto mode does the same when no AI Gateway credentials are set. The ai-gateway mode never reads native subscriptions. Native subscription credentials stay on the host. Just like with API keys, credentials are resolved at the host boundary. OAuth access tokens are refreshed there as needed. Where the sandbox supports it, the harness receives placeholder credentials and the real token is injected into outbound requests on the host. Native subscriptions work across harness adapters wherever the harness itself supports subscription login. These include Claude Code, Cline, Codex, Cursor, fx, GitHub Copilot, Grok Build, OpenCode, and Pi. Read the harness documentation to learn more. Read more ## Manage end-user OAuth consent for AI agents with Amazon Bedrock AgentCore DevFeed: [Manage end-user OAuth consent for AI agents with Amazon Bedrock AgentCore]() Original publisher: [Read original article]() Author: Swara Gandhi Published: 2026-09-14T20:35:45Z Content type: tutorial Language: en Sources: [Artificial Intelligence]() Topics: [Amazon Bedrock AgentCore](), [OAuth](), [Amazon Web Services](), [AWS CloudTrail](), [GitHub](), [Slack](), [Model Context Protocol](), [ide]() Tags: [amazon-bedrock](), [amazon-bedrock-agentcore](), [announcements](), [aws](), [aws-cloudtrail](), [claude-code](), [cursor](), [github](), [intermediate-200](), [kiro](), [model-context-protocol](), [oauth](), [slack](), [visual-studio-code]() ### AI overview This tutorial explains how Amazon Bedrock AgentCore Identity's Consent portal manages end-user OAuth consent and session binding for AI agents. It covers configuring GitHub and Slack targets through an AgentCore Gateway, storing user tokens, supporting IDE and MCP clients, and reviewing activity in AWS CloudTrail. ### Source excerpt Amazon Bedrock AgentCore Identity now offers a Consent portal, a managed web experience and session binding endpoint for AgentCore Gateway. This post walks through provisioning a portal, configuring GitHub and Slack 3LO targets, and the end-user consent flow, and shows how to review activity in AWS CloudTrail. ## Your AI coding spend bought 25% more output. Duplication rose 81%. DevFeed: [Your AI coding spend bought 25% more output. Duplication rose 81%.]() Original publisher: [Read original article]() Author: Steve Fenton Published: 2026-09-14T14:39:14Z Content type: article Language: en Sources: [The New Stack]() Topics: [ai-coding](), [Claude Code](), [cursor](), [datasets]() Tags: [agentic](), [ai](), [ai-coding](), [ai-impact](), [ai-operations](), [ai-strategy](), [ai-tools](), [claude-code](), [coding](), [contributed](), [contributed-octopus-deploy](), [cursor](), [dataset](), [software-development]() ### AI overview The article examines the return on investment from AI coding tools. It reports that heavy users gained 25% over their previous velocity, while code duplication rose 81%, and argues that output measures such as lines of code, pull requests, and feature counts do not reliably represent business value. ### Source excerpt Since they arrived on the scene, a great swathe of the software industry has pinned its hopes on AI tools, The post Your AI coding spend bought 25% more output. Duplication rose 81%. appeared first on The New Stack. ## "Same mission, bigger stage": OpenAI hires Git AI founders to help Codex prove its ROI DevFeed: ["Same mission, bigger stage": OpenAI hires Git AI founders to help Codex prove its ROI]() Original publisher: [Read original article]() Author: Paul Sawers Published: 2026-09-12T14:46:53Z Content type: news Language: en Sources: [The New Stack]() Topics: [ai observability](), [cursor]() Tags: [ai](), [ai-agents](), [ai-coding](), [codex](), [cost](), [cursor](), [developer-tools](), [git](), [open-source](), [openai]() ### AI overview OpenAI hired Git AI's founders to help measure the performance, cost, and return on investment of Codex and other AI coding tools. ### Source excerpt OpenAI has hired the founders of Git AI, an open-source tool that tracks how much code is written by AI The post "Same mission, bigger stage": OpenAI hires Git AI founders to help Codex prove its ROI appeared first on The New Stack. ## Test Complex Interactions Earlier with AI Prototyping DevFeed: [Test Complex Interactions Earlier with AI Prototyping]() Original publisher: [Read original article]() Author: Megan Chan Published: 2026-09-11T17:00:00Z Content type: article Language: en Sources: [NN/g latest articles and announcements]() Topics: [Artificial Intelligence](), [User interface design](), [cursor]() Tags: [ai](), [article](), [case-studies](), [conversational-ai](), [cursor](), [design](), [prototyping]() ### AI overview AI prototyping tools enable teams to create high-fidelity, interactive prototypes of complex interfaces and test them with users earlier in the design process. The article discusses how Ramp used Cursor to prototype an AI chat-based expense-policy editor and uncover usability issues and edge cases. ### Source excerpt AI tools make it feasible to build fully interactive prototypes of complex interfaces so you can test them with users earlier in the design process. ## GitHub Copilot is now available in the AI SDK harness layer DevFeed: [GitHub Copilot is now available in the AI SDK harness layer]() Original publisher: [Read original article]() Author: Felix Arntz Published: 2026-09-10T17:39:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [GitHub Copilot](), [vercel ai sdk](), [AI-assisted coding](), [cursor]() Tags: [agent](), [ai](), [claude](), [claude-code](), [code](), [codex](), [coding](), [copilot](), [cursor](), [github](), [github-copilot]() ### AI overview The AI SDK harness layer adds official GitHub Copilot support through an adapter, allowing applications to run Copilot and other coding agents through the same HarnessAgent interface. ### Source excerpt The AI SDK harness layer now supports GitHub Copilot through the official @ai-sdk/harness-github-copilot adapter. The harness layer lets your application run different coding agents through the same HarnessAgent interface, so you can switch agents without changing your application code. Pass githubCopilot to HarnessAgent: Under the hood, the adapter uses @ai-sdk/harness-acp to connect GitHub Copilot to HarnessAgent through the Agent Client Protocol (ACP). Supported harnesses now include, in addition to GitHub Copilot, Claude Code, Cline, Codex, Cursor, Deep Agents, fx, Grok Build, OpenCode, and Pi, with more coming soon. Read the GitHub Copilot harness documentation to get started. Read more ## Your tools work. Will the agent use them right? DevFeed: [Your tools work. Will the agent use them right?]() Original publisher: [Read original article]() Author: Gil Levin Published: 2026-09-10T00:00:00Z Content type: article Language: en Sources: [Webflow Blog]() Topics: [Model Context Protocol (MCP)](), [Artificial Intelligence](), [Claude Code](), [cursor]() Tags: [ai-agents](), [chatgpt](), [claude-code](), [cursor](), [engineering](), [mcp](), [tools]() ### AI overview Webflow's MCP eval harness tests whether autonomous agents use its tools correctly to complete plain-English, multi-step tasks. It runs Claude Code or OpenAI's Codex CLI against disposable Webflow sites, combines deterministic checks with LLM and visual judging, and sends traces and scores to Datadog. ### Source excerpt A passing MCP tool test tells you the JSON is valid. It doesn't tell you an agent will pick the right tool, in order, and get the task right. ## How cloud native goes AI native DevFeed: [How cloud native goes AI native]() Original publisher: [Read original article]() Author: Doron Grinstein, CEO of Control Plane Published: 2026-09-09T08:13:23Z Content type: opinion Language: en Sources: [Cloud Native Computing Foundation]() Topics: [cloud-infrastructure](), [App](), [coding](), [site-reliability-engineering](), [Database](), [Claude](), [cursor]() Tags: [ai](), [apps](), [blog](), [claude](), [cloud](), [cursor](), [infrastructure](), [production](), [sre]() ### AI overview The article argues that AI-assisted, "vibe-coded" applications are increasingly easy to start but often fail to reach production because they bypass cloud-native operational practices. It frames closing that production-readiness gap as a key infrastructure challenge. ### Source excerpt "A sales guy writing code" used to be the lead-up to a joke. But now no one's laughing. Designers used to sit meekly waiting for the high priests of code to make their designs real. Now... ## DeepSeek V4.1 Flash now available on AI Gateway DevFeed: [DeepSeek V4.1 Flash now available on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-09-09T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [deepseek](), [Artificial Intelligence](), [API](), [Language models](), [AI-assisted coding](), [Claude Code](), [codex](), [cursor]() Tags: [ai](), [ai-gateway](), [api](), [caching](), [codex](), [context-window](), [cursor](), [deepseek](), [responses](), [tool](), [vercel](), [vision]() ### AI overview DeepSeek V4.1 Flash is now available through Vercel's AI Gateway, offering native image understanding, a 1 million token context window, responses up to 384,000 tokens, reasoning, tool use, and prompt caching. Developers can use it in Claude Code, Codex, Cursor, and other coding agents with the model name deepseek/deepseek-v4.1-flash. ### Source excerpt DeepSeek V4.1 Flash is now available on AI Gateway with native image understanding. V4.1 Flash has vision support and accepts text and images in the same request, so you can ask questions about screenshots, read charts, and extract information from visual content. The model has a 1 million token context window and supports responses up to 384,000 tokens, along with reasoning, tool use, and prompt caching. Its new architecture processes input and generates output with separate components, reducing the active computation needed for each stage. Use deepseek/deepseek-v4.1-flash as the model name: To use it in Claude Code, Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select deepseek/deepseek-v4.1-flash in the agent. See the coding agents guide for details. Try DeepSeek V4.1 Flash in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. You can view all language models available on AI Gateway. Read more ## Bring Claude and Cursor Coding Agents Into Your Port Catalog DevFeed: [Bring Claude and Cursor Coding Agents Into Your Port Catalog]() Original publisher: [Read original article]() Author: Sagi Keren Published: 2026-09-07T13:53:22Z Content type: article Language: en Sources: [Developer Experience & Platform Engineering Blog | Port]() Topics: [cursor](), [coding](), [Claude](), [AI-assisted coding](), [Orchestration](), [Automation]() Tags: [agents](), [ai](), [automation](), [claude](), [cloud](), [coding-agents](), [cursor](), [governance](), [workflows]() ### AI overview The article explains how Port integrates Claude and Cursor coding agents into a centralized catalog. It focuses on governance, ownership, auditing, workflow integration, and self-service creation through governed templates. ### Source excerpt Port catalogs, creates, and triggers Claude and Cursor coding agents, giving platform teams one governed registry for AI agents. ## The iOS Simulator Inside Cursor, Claude Code, and Codex DevFeed: [The iOS Simulator Inside Cursor, Claude Code, and Codex]() Original publisher: [Read original article]() Author: Antoine van der Lee Published: 2026-09-07T13:07:15Z Content type: opinion Language: en Sources: [SwiftLee]() Topics: [iOS](), [cursor](), [Claude Code](), [codex](), [App](), [Tooling]() Tags: [agent](), [agent-skills](), [ai-development](), [app-development](), [claude-code](), [codex](), [cursor](), [developer-tools](), [ios]() ### AI overview An opinion article examines how the iOS Simulator is becoming integrated into Cursor, Claude Code, and Codex as app development shifts toward orchestrating AI agents. It argues that visual testing, performance review, and critical human validation remain essential, and describes AI Agents for the iOS Simulator as a way to automate navigation and validation within the developer workflow. ### Source excerpt App development is changing, and the iOS Simulator is too, as it starts to appear in Cursor, Claude Code, and Codex in different ways. I've been developing apps since 2009 and the RocketSim developer tool since 2019, and I can tell we're reinventing ourselves like never before. I also started to wonder: what is the ... -> The post The iOS Simulator Inside Cursor, Claude Code, and Codex appeared first on SwiftLee. ## How to add SSO to your B2B SaaS app with WorkOS AuthKit DevFeed: [How to add SSO to your B2B SaaS app with WorkOS AuthKit]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-09-07T00:00:00Z Content type: tutorial Language: en Sources: [WorkOS Blog]() Topics: [Single sign-on (SSO)](), [Software as a service](), [Authentication](), [SDKs]() Tags: [asp-net](), [asp-net-core](), [authentication](), [b2b](), [claude](), [cli](), [codex](), [configuration](), [cursor](), [how-to](), [saas](), [sdk](), [sso]() ### AI overview A tutorial on adding SSO to a B2B SaaS application with WorkOS AuthKit. It explains that AuthKit handles SSO-specific SAML operations, while developers must design the organization model, configure who manages connections, account for existing users, and test failure cases. ### Source excerpt Enabling SSO is a dashboard toggle. The real work is the organization model, self-service configuration, and the traps that surface after your first enterprise customer. ## YOLO Mode: Agent Autonomy Without the Guardrails DevFeed: [YOLO Mode: Agent Autonomy Without the Guardrails]() Original publisher: [Read original article]() Author: Srini Sekaran Published: 2026-09-03T18:00:00Z Content type: article Language: en Sources: [Docker]() Topics: [AI-assisted coding](), [Authorization](), [Command-line interface](), [cursor](), [GitHub Copilot CLI](), [Claude]() Tags: [agentic-ai](), [ai-agent](), [cli](), [concepts](), [cursor](), [docker-ai-governance](), [docker-sandboxes](), [products](), [sandbox](), [secrets](), [security]() ### AI overview The article defines YOLO mode as allowing AI coding agents to auto-approve actions without confirmation prompts. It argues that the key risk depends on the execution environment and recommends isolated, disposable sandboxes with scoped access and no real secrets. ### Source excerpt YOLO mode lets an AI agent run without asking permission. Learn what it is, why it's risky, and how to run it safely. ## Cursor Cloud Agents can now run in Vercel Sandbox DevFeed: [Cursor Cloud Agents can now run in Vercel Sandbox]() Original publisher: [Read original article]() Author: Allen Zhou Published: 2026-09-03T15:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [cursor](), [Vercel](), [Agent Harness](), [Firecracker]() Tags: [agents](), [apis](), [architecture](), [cursor](), [enterprise](), [firecracker](), [guide](), [sandbox](), [scale](), [vercel](), [virtual-machines](), [workflows]() ### AI overview Cursor Cloud Agents can run in Vercel Sandbox, using isolated Firecracker microVMs as their execution environment. The architecture combines Cursor's agent harness and inference loop with Vercel Functions and Workflows for durable request handling, worker provisioning, monitoring, retries, cleanup, and short-lived user-scoped credentials. ### Source excerpt Cursor Cloud Agents can now run in Vercel Sandbox instead of Cursor's hosted machines. Cursor manages the agent harness and inference loop. Its Self-Hosted Machines APIs let you supply the execution environment where agents clone repositories, edit files, and run commands and tests. Self-Hosted Machines requires a Cursor Enterprise plan. Vercel Sandbox provides that execution environment as an isolated Firecracker microVM for each agent request. Vercel Functions and Vercel Workflows form a durable control plane that claims queued agent requests, provisions workers, monitors sessions, and cleans up automatically. With this architecture, you get: A scale-to-zero worker pool without long-lived virtual machines A dedicated, isolated Sandbox for every agent request Durable retries when a worker or session fails Short-lived, user-scoped credentials inside each Sandbox Follow the step-by-step guide to deploy the reference implementation to your own Vercel account, or learn more in the Vercel Sandbox documentation. Read more ## 5 ways to optimize AI costs and reduce wasted AI spend DevFeed: [5 ways to optimize AI costs and reduce wasted AI spend]() Original publisher: [Read original article]() Author: info@1password.com (Rachel Sudbeck) Published: 2026-09-03T00:00:00Z Content type: article Language: en Sources: [Blog on 1Password Blog]() Topics: [Artificial Intelligence](), [cursor]() Tags: [ai](), [ai-agents](), [ai-models](), [cost](), [cursor](), [saas-management](), [shadow-ai](), [unified-access]() ### AI overview The article outlines ways businesses can reduce AI spending, including choosing less expensive models, overseeing AI agents, and identifying shadow AI. ### Source excerpt The tokenmaxxing era has left companies grappling with an uncomfortable reality. Now that AI vendors have switched to usage-based billing models, businesses are facing sky-high bills, and IT and finance teams are under pressure to rein in spending without slowing down innovation. The logical first step is to locate areas where that spend is going to waste, but even getting visibility into usage can be overwhelming when it's spread across departments, users, models, vendors, and agents. If you're trying to track down wasted AI spend and find opportunities to optimize your tokens, it helps to start with some of the primary reasons why AI bills may balloon past your company's budget. Top ways to optimize your company's AI costs So IT and Finance teams can know where to focus their efforts, here are five of the most common sources of unexpected AI spend. 1. Stop defaulting to the most expensive model For businesses to optimize spend, they need a way of overseeing and enforcing which models are being used for what tasks. Different AI models can vary wildly both in their abilities and their cost, and many users default to flagship AI models without realizing that there are more affordable options that can accomplish their goals at a fraction of the cost. For instance, in a recent experiment run by Cursor, building a web browser from scratch cost $10,565 when using a top-tier flagship model, and $1,339 when using a mix of models, even though the end results were comparable in terms of quality. 2. Stop letting agents run without oversight As the Stanford Digital Economy Lab reported, AI agents are "uniquely expensive, consuming 1000x more tokens than code reasoning and code chat." Meanwhile, data from OpenRouter shows that the majority of tokens spent overall are being used by agents. Here's a scenario that's becoming familiar to many AI developers and builders: An agent is instructed to perform a certain task, but it fails. So it tries again, and fails. With each loop, it ## Build and run Datadog workflows from Bits Chat or AI agents DevFeed: [Build and run Datadog workflows from Bits Chat or AI agents]() Original publisher: [Read original article]() Author: Neha Haresh; Gabriel Margolis; Brianna Wang; David Robert-Ansart Published: 2026-09-03T00:00:00Z Content type: tutorial Language: en Sources: [Datadog | The Monitor blog]() Topics: [AI Bots](), [cursor](), [incident]() Tags: [ai](), [ai-agents](), [ai-coding](), [bits-ai](), [claude-code](), [codex](), [cursor](), [debug](), [incident-management](), [mcp](), [workflow-automation](), [workflows]() ### AI overview Datadog describes using Bits Chat and AI coding agents to create, run, and debug automated workflows with Datadog and development context. ### Source excerpt Build, debug, and run Datadog workflows from Bits Chat and AI coding agents using the operational and development context where you're working. ## Your Agent Speaks MCP. Give It a Computer. DevFeed: [Your Agent Speaks MCP. Give It a Computer.]() Original publisher: [Read original article]() Published: 2026-09-03T00:00:00Z Content type: article Language: en Sources: [The Fly Blog]() Topics: [MSP MCP](), [AI Bots]() Tags: [agents](), [api](), [auth](), [cdn](), [claude](), [claude-code](), [cli](), [close-to-users](), [cloud](), [codex](), [cursor](), [deploy-app-servers](), [docker](), [elixir](), [filesystems](), [fly](), [fly-io](), [gemini](), [heroku-alternative](), [heroku-competitor](), [hosting](), [i](), [mcp](), [mcp-server](), [networking](), [postgresql-clusters](), [servers](), [skills]() ### AI overview The article presents Sprites as disposable cloud computers for coding agents and explains using their API through MCP. It argues that MCP transport and progressive capability disclosure can work together, including through plugins and skills. ### Source excerpt Sprites are disposable cloud computers. They appear instantly, always include durable filesystems, and cost practically nothing when idle. They're the best and safest place on the Internet to run agents and we want you to create dozens of them. Sprites are a place to run agents; the first thing you should think to do with a new Sprite is to type claude (or gemini or codex). We've put a lot of effort into making sure coding agents feel safe and happy when they're on Sprites, because, to (probably) quote John von Neumann, "happy agents are productive agents." What's less obvious about Sprites is that they're great tools for agents. Want three different versions of a new feature? A test environment? An ensemble of cooperating services? It's super handy to be able to start your prompts, "On a new Sprite, do...". The Sprites API is simple, discoverable, and designed for this use case. The only real question is how your agent reaches it. For most of you the answer is MCP, and the setup is already written. You Don't Have To Pick There's an argument going around that MCP is the wrong way to extend an agent, and that command line tools and discoverable APIs are the Right Way. Half of that argument is correct, and it's the important half, so let's take it seriously. Dumping thirty tool descriptions into a context window is a bad way to teach anything. Not every Sprite command matters in every session, and cramming them all in signals to the model that they all matter to you. If you're not using network policies, gemini shouldn't burn a single token learning to configure them. Capabilities should reveal themselves progressively, the way they do when an agent works out a CLI one subcommand at a time. The wrong half is treating that as a case against MCP. Progressive disclosure is a question of what you say to the model. MCP is a question of how the bytes get there: transport, auth, structured results, a tool the model can call instead of a command whose flags it has to guess. Tho ## Muse Spark 1.3 now available on AI Gateway DevFeed: [Muse Spark 1.3 now available on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-09-02T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [AI Chat]() Tags: [agents](), [ai](), [ai-gateway](), [claude](), [claude-code](), [codex](), [coding](), [cursor](), [meta](), [models](), [pdf](), [playground](), [release](), [spark](), [vercel]() ### AI overview Muse Spark 1.3 from Meta is now available on AI Gateway in standard and contributor pricing tiers. The model adds a 1M-token context window and accepts text, image, and PDF input; the contributor tier uses the same model but allows Meta to use submitted inputs and outputs for model improvement in exchange for lower pricing. ### Source excerpt Muse Spark 1.3 from Meta is now available on AI Gateway, in both the standard and contributor pricing tiers. This model improves on prior Muse Spark models at agent work and coding, with a 1M token context window and text, image, and PDF input. On coding it takes fewer turns and writes less filler than the previous release. To use Muse Spark 1.3, set model to meta/muse-spark-1.3: Contributor tier Muse Spark 1.3 Contributor is a pricing tier on the same model rather than a separate one, with the same weights, capabilities, and context window. The difference is that Meta uses the inputs and outputs sent to this tier to train and improve its models, and pricing is lower in exchange. Model Input Output Cached input meta/muse-spark-1.3 $1.25 $4.25 $0.15 meta/muse-spark-1.3-contributor $0.10 $0.20 $0.002 Rates are per million tokens and unchanged from Muse Spark 1.2 on both tiers. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, Cursor, and more, then select meta/muse-spark-1.3 in the agent. Try Muse Spark 1.3 in the model playground. You can view all language models available on AI Gateway. Read more ## GLM-5.3 is 50% off through DigitalOcean on AI Gateway DevFeed: [GLM-5.3 is 50% off through DigitalOcean on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-09-02T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [Vercel](), [Digital Ocean](), [Language models](), [long-context](), [AI-assisted coding](), [Claude Code](), [codex](), [cursor]() Tags: [ai-gateway](), [claude-code](), [codex](), [coding-agents](), [context-window](), [cursor](), [digitalocean](), [language-models](), [model](), [tokens](), [vercel]() ### AI overview Vercel announces a 50% discount on GLM-5.3 through DigitalOcean on AI Gateway until September 8. The article explains the temporary promo model name, standard provider routing, model limits, spend tracking, and setup for coding agents. ### Source excerpt GLM-5.3 is 50% off on AI Gateway through Tuesday, September 8, in partnership with DigitalOcean. How to use the model during the offer period Using the promo name (zai/glm-5.3-promo-50) gets the discounted rate. It routes only to DigitalOcean, with no fallback to another provider, and it stops serving when the offer ends. Using the standard name (i.e., zai/glm-5.3) with provider options to sort DigitalOcean as the preferred provider keeps working after September 8 and routes across every provider that serves the model, at their usual rates. Because the promo name goes away when the offer ends, treat it as something you switch on for the window rather than hardcode. To keep the standard name in your code instead, pin the provider with order: ['digitalocean'] under providerOptions.gateway, which prefers DigitalOcean and falls back to the others if it cannot serve the request. GLM-5.3 takes text input, with a 1M token context window and a maximum output of 128K tokens. Discounted requests appear in your spend dashboard and carry a trace like any other request. Try GLM-5.3 in the model playground. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, Pi, and more and select zai/glm-5.3-promo-50 inside the agent. You can view all language models available on AI Gateway. Read more ## Encore now works with Entire DevFeed: [Encore now works with Entire]() Original publisher: [Read original article]() Author: Ivan Cernja Published: 2026-09-02T00:00:00Z Content type: tutorial Language: en Sources: [Encore Updates]() Topics: [AI-assisted coding](), [codex](), [Claude Code](), [coding](), [Git](), [cursor](), [PostgreSQL]() Tags: [claude-code](), [codex](), [coding](), [coding-agents](), [cursor](), [git](), [object-storage](), [postgresql]() ### AI overview Encore now works with Entire, allowing developers to use Codex, Claude Code, Gemini, and Cursor in Encore projects while recording agent sessions in Git and linking them to commits. The article demonstrates building and testing an image service with local PostgreSQL and object storage, including integration tests and a live upload/download verification. ### Source excerpt Keep the agent session behind every Encore change, directly in Git. ## Mobile UI Testing with Claude Code, Cursor and Codex DevFeed: [Mobile UI Testing with Claude Code, Cursor and Codex]() Original publisher: [Read original article]() Author: Ashish Kharche Published: 2026-09-01T10:00:00Z Content type: tutorial Language: en Sources: [mobile.dev - Medium]() Topics: [Testing](), [Mobile](), [Model Context Protocol](), [Android](), [iOS](), [Emulator](), [simulator]() Tags: [agent](), [android](), [chromium](), [claude-code](), [cli](), [codex](), [cursor](), [emulator](), [ios](), [maestro](), [mcp](), [mobile](), [regression](), [simulator](), [testing](), [yaml]() ### AI overview This tutorial explains how Claude Code, Cursor, Codex, and other MCP-enabled coding agents can use Maestro MCP to explore mobile app interfaces on iOS Simulators, Android emulators, physical Android devices, and Chromium, then save the tested steps as reusable Flow files. The saved Flows can be rerun through an agent or the CLI, including in CI, to detect UI changes. ### Source excerpt Have Claude Code, Cursor or Codex run a UI test on an iOS Simulator or Android emulator, then save it as a Flow and run the same test again. ## Agents babysitting agents DevFeed: [Agents babysitting agents]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-08-31T18:26:02Z Content type: opinion Language: en Sources: [WorkOS Blog]() Topics: [Development](), [AI-assisted coding](), [Pull Request](), [Claude Code](), [Model Context Protocol (MCP)](), [Terminal](), [Monitoring](), [cursor]() Tags: [agent](), [agents](), [claude](), [claude-code](), [coding-agents](), [cursor](), [mcp](), [monitoring](), [pull-request](), [review](), [terminal](), [tests]() ### AI overview WorkOS engineers describe using one coding agent to supervise another after pull requests are opened, with headless monitoring, narrowly scoped jobs, visible costs, protected credentials, and human approval for merging. ### Source excerpt How WorkOS engineers make agents babysitting agents safe: a headless surface to supervise through, scoped jobs, visible costs, and a human on the merge button. [Next page]()