#MCP
50 articles with this tag
Most read in #MCP, last 30 days
LinearB sells context as the fix for AI code
LinearB launched Agentic Studio to turn delivery data into plain language answers for leaders and context for coding agents, with agent skills in early access.
Vercel Wants One Person And Agents To Build $1T
Vercel CEO Guillermo Rauch told Delphi his platform could host a single-person trillion-dollar company, with AI agents doing the work of thousands.
Oracle Shows Single MCP Gate to Lock Agent Data
Oracle's Jeff Smith showed on AI Engineer how a single MCP server with governed reports and deep data security can give agents data without leaking PII.
Who Made That Request? Agent Auth Has No Opt-Out
Kamil Potrec told AI Engineer that agents need their own identities and short-lived, scoped credentials, with gateways to keep tokens out of agent reach.
Stop handing tools to agents, give them sandboxes
Arize AI's Mikyo King told AI Engineer agents should run code in sandboxes, not juggle tool catalogs, cutting tokens 43% but raising new isolation risks.
Agents Need Bosses, Badges and Boundaries
Tim Tyler told AI Engineer that governing agents like software misses the point, proposing badges, sponsors and scorecards instead.
Runlayer says it tracked 300 agents for 60 staff
Runlayer found 300 daily agents for 60 staff and detailed how remote prompts and over-permissioned tools enable exfiltration.
Google says managed agents push code off your laptop
Jack Wotherspoon demoed Gemini managed agents that run the anti-gravity harness in a cloud sandbox and keep secrets behind a proxy.
Agents escaped the sandbox. Identity is the fix.
Andrey Rybka detailed real agent breaches at AI Engineer, from sandbox escapes to DNS theft, and mapped fixes to identity and deterministic controls.
Stop handing agents your user token
WorkOS demoed off.md and ID-JAG to let AI agents sign up safely, plus scoped agent tokens that limit what agents can do.
Stripe Lets Agents File Your Startup, With Limits
Stripe now lets you start Atlas incorporation from an AI agent or CLI, but signatures and payments stay in the Dashboard.
Cloudflare turned on WebMCP for 100,000 stores
Cloudflare says it enabled WebMCP at the edge for almost 100,000 Commerce.com storefronts to let shopping agents call tools instead of scraping HTML.
Vanguard Cut Research From 7 Days to 15 Minutes
Vanguard's Gareth Yoder said exposing legacy code via MCP cut research from 7 days to 15 minutes during a cash ledger modernization handling 30M calls a day.
Stripe hides your card from agents, keeps approval
Stripe's link-cli lets agents request one-time virtual cards and tokens, but every spend still needs a push approval and the MCP server stays loopback-only.
Anthropic split its agent to cut start time 90%
Anthropic said decoupling the harness from the sandbox cut time to first token by 90% and moved credentials out of execution to block prompt injection.
Flywire's Orbit bets on guardrailed workflows
Flywire's Orbit ran 11,000 workflows at 99% success with guardrailed YAML workflows, but sandboxing and online evaluations remain incomplete.
AWS Let 1,000 Agents Attack Its Own Sandbox
An AWS engineer let coding agents run 1,000 sandbox tasks overnight and mapped where MCP and container primitives break them.
Oracle says agents forget, and that is a risk
Oracle's Anders Swanson told AI Engineer that agent context is not memory and demoed a governed, hybrid retrieval layer to stop poisoned and stale recall.
Airbyte Built Its CLI for Agents, Not People
Airbyte engineer Pedro Lopez told AI Engineer why agents need JSON, uniform commands and skills, and why MCP authentication still breaks across clients.
Web Search APIs for AI Agents: Benchmarks, Pricing and MCP Setup
Eleven web search APIs for autonomous agents compared on latency, cost per 1,000 calls and factual accuracy, with the MCP wiring for Claude Code.
How to Find Stealth Startups in 2026
Learn how to find stealth startups on StartupHub in 2026. Filter search, open a profile, check a known domain, set a monitor and email the company.
Snowflake wants agents to run your pipelines alone
Snowflake's 5-stage autonomous data engineering model promises zero overhead but leaves security and governance proof for later.
Sam Altman wrong about AI: $700B trap
Eli the Computer Guy says OpenAI's bigger-is-better bet locked in $700B in contracts as enterprises shift to bespoke, fine-tuned models.
Shu Fang Two Sigma Tethered agents
Two Sigma runs cloud agents as the user via Kubernetes sidecars, tagging with trace headers and VPC-bound Google grounding.
Databricks Unity Gateway Cut $1.2M Agent Waste
Databricks found seven MCP bugs wasting $1.2M a year and fixed them in an hour using Unity Gateway traces and Genie One.
The Lethal Trifecta Blocking Agentic Commerce
David Levine says the lethal trifecta, private data plus untrusted content plus action, blocks agentic commerce until agent identity is verifiable.
Generative UX Ends Static Screens at Scale
Commercetools showed intent-driven UI that renders live components from prompts, and why declarative protocols need strict curation to be safe.
Agent Spending Without Controls: Demo Shows Risk
Edge & Node showed how Ampersend adds a compliance harness to x402 so AI agents can pay for tools but get blocked if wallets are sanctioned.
x402 Isn't Good Yet, Apify Says
Apify 10x'd the x402 market to 20,000 tools, but CEO Jan Curn says double spending, metering and 402 vs 401 conflicts keep it from production.
DeepMind's Multimodal Collaborative Agents
Google DeepMind's Nidhi Kaushik Vyas demos a three-phase loop for agents that handle fuzzy shopping intent with visual elicitation.
Databricks Genie Ontology Needs Clean Data First
Databricks says Genie Ontology needs six layers of data work to deliver accurate answers, starting with clean facts and golden records.
AI Agents Are Where Microservices Were in 2015
Navan's architects map the emerging agentic stack and argue runtime and memory are solved, while observability, testing and cost control are not.
ChatGPT Work Turns Transcripts Into Templates
ChatGPT Work converts meeting transcripts into templated Google Docs and Slides, then reuses the formatting for every future update.
AI Native Organizations Run on Skills
QuantumBlack's Imad Touil says AI-native scale depends on governed, portable skills, not just models and tools.
Figma's Coding Agents Playbook Without Garbage
Figma engineer Alon Blum shares how verification, planning and attention-aware docs let coding agents ship 5x faster without trashing quality.
Frontier Development Hits 4.5x Gains at Amazon
Amazon's Claire La Gory details how frontier development delivers 4.5x median gains with Kiro when teams change habits, not just tools.
Claude's Corner: Corvera - AI Agents for the CPG Operations Problem
Corvera deploys AI agents to automate the back-office grind that kills CPG brands at scale. Here's how the YC W2026 startup built an agentic OS on MCP, why domain expertise is the real moat, and a full build guide for replicating it.
Snowflake Tightens AI Agent Controls
Snowflake enhances its CoCo AI agent platform with new governance features for cost, access, and tool usage.
Cloudflare adds WriteGuard for AI agent safety
Cloudflare introduces WriteGuard, a new feature for its MCP server portals, offering fine-grained controls and auditing for AI agents to prevent misuse.
MCP Tasks: Why Agents Aren't Supporting Them
Cornelia Davis of Temporal explains why MCP tasks are complex and why adoption is slow, previewing V2 improvements.
Snyk Tackles Agentic Development Security
Snyk's Ezra Tanzer discusses the critical security challenges in agentic development, focusing on securing agent outputs, supply chains, and behavior.
Skills Are the New SDKs: Rethinking AI Agents
Elvin Aghammadzada of DataRobot argues that 'skills' are the new SDKs for AI agents, addressing context engineering challenges and the shift from 'friction' to 'fluency' moats.
Pomerium's Nick Taylor Enhances OpenClaw Security and UX
Nick Taylor, Developer Advocate at Pomerium, introduced a new "trusted proxy auth mode" for OpenClaw, enhancing security and user experience by eliminating manual token management and showcasing live AI application development.
Pietro Zullo on MCP Apps and the Future of Software
Pietro Zullo of Manufact, Inc. explains MCP Apps, a new framework for building AI-powered software with primitives like `ui/update-model-context` and `ui/message`.
MCP & ChatGPT Apps: The Double Iframe Strategy
Frédéric Barthelet of Alpic explains how MCP and ChatGPT apps use double iframes and CSP for secure, interactive UI embedding within chat interfaces.
Claude's Corner: Orthogonal, The API Economy, Rebuilt for Agents
Orthogonal gives AI agents instant access to 50+ premium APIs through a single MCP server and pay-per-call credits. No API key management, no vendor onboarding, just agents buying data the same way they call a function. Here's how it works and whether you can replicate it.
Robinhood Lets AI Trade Your Stocks
Robinhood's new Agentic Trading feature lets AI agents manage investments, using the Model Context Protocol (MCP) for seamless integration and automated trades.
Rachel Nabors: The Infinite Canvas of the Web Agent
Rachel Nabors discusses how AI agents can leverage the web's 'infinite canvas' using MCP tools and WebMCP, transforming browser interactions.
Uber Tackles AI Agent Identity
Uber is enhancing AI security with a new identity system for autonomous agents, ensuring accountability and traceability in complex workflows.
Claude Agents Get Private Sandbox
Anthropic's Claude Managed Agents now offer self-hosted sandboxes and MCP tunnels, bringing agent execution and private service access under enterprise control.
