A developer has released an open-source WhatsApp MCP server that runs in the user's own Cloudflare account. After linking a phone the way WhatsApp Web is linked, any MCP client such as Claude Code or Claude Desktop can search and read chats but cannot send messages, with messages stored in a SQLite database inside a Durable Object. The developer says the tool is unofficial, uses the WhatsApp Web protocol through the open-source Baileys library in violation of WhatsApp's terms, and has only been tested against one real account; importing large histories requires a paid Workers plan.
The Verge reports that users of agents including Meta's Muse, xAI's Grok, OpenAI's Dots and Instinct have in recent weeks hit blocks preventing them from completing tasks on major websites. Yelp said it only allows AI agents that pay for its data licensing program, Delta Air Lines is stopping agents from booking tickets on users' behalf, and JPMorgan Chase and Charles Schwab confirmed they block agent access to accounts over security concerns; users also report Cars.com and Amtrak no longer working. The piece recounts Amazon's legal fight with Perplexity over agentic shopping and notes Meta has instead partnered with retailers such as Walmart, Gap, Sephora and Best Buy, while Shopify has opened its storefront network to AI agents.
A commentary republished by Huxiu AI from Token Moment author Cassie argues that as personal assistants such as Muse, Instinct, Grok Bot, WeChat Xiaowei and Cue appear, users are handing search, browsing, price comparison, ordering and payment to AI agents, while Amazon has blocked Muse's access and Apple has tightened agent permissions. Citing Cloudflare, a16z and Alipay figures, the piece says AI traffic share and agent token consumption are rising fast, and outlines three platform responses: competing for the human entry point, exposing back-end capabilities as agent-callable tools, or shifting to scarce personalized experiences. It also mentions Meta and Sierra working with Shopify, Stripe, Walmart and others on a Personal Agent Protocol, though no first-hand sources are given for the statistics or for the blocking and permission details.
OpenAI and other labs' AI models have solved decades-old math problems, prompting cryptography experts to debate whether powerful AI could bypass private keys and break encryption. Whitfield Diffie, a co-inventor of public-key cryptography, said he expects major changes in mathematics and cryptanalysis but cannot gauge the timeline — a year may be too soon and five years may not be enough. Martin Hellman is more optimistic, while Johns Hopkins professor Matthew Green said public-key cryptography may be obsolete and predicted frontier AI could fundamentally break elliptic-curve cryptography within two years; Cloudflare's 2025 report says X25519 is used by most internet traffic. A Princeton computer science professor argued well-studied mainstream algorithms will likely stay safe, but less-studied ones, including many post-quantum candidates, could be broken or weakened; in July Anthropic said its AI nearly broke a NIST post-quantum candidate algorithm and had briefed the US government and industry partners beforehand. OpenAI did not respond on whether its internal models were used against cryptography.
In a panel hosted by The Information co-executive editor Amir Efrati, executives from Block, Cloudflare and Klaviyo discussed how companies should be organized when AI agents can understand context, coordinate work and act across systems. Cloudflare chief strategy officer Stephanie Cohen said the company had to redesign the work itself, including eliminating whole functions, while Block's Owen Jennings said Block rebuilt its functional model so engineering, product, design and sales work across Square, Cash App and Afterpay on shared capabilities, and open-sourced Buzz, a workspace giving each agent its own identity, permissions and context. Jennings said Block shipped about three times as many Square features in the first half of 2026 as a year earlier — a company-reported figure — and Klaviyo CTO Surabhi Gupta said Klaviyo tracks development velocity, cost and hours saved; no panelist would predict headcount.
Writing up Theory's second annual AI in Practice Survey, fielded on 22 September 2026 with 455 senior technical builders, Tomasz Tunguz reports that security and governance rose from sixth to first among pain points in twelve months. 43% of respondents rate security and governance the most important pain point, with evals a distant second at 22%; only 10% say security tooling delivers the most value today, while 18% expect it to lead in future. Cost optimisation and model selection are now standard: 45% of teams route queries across models (23.7% on a router they trained, 21.5% on a proprietary one) while 54.7% send every query to a single model. Among the 259 teams whose agents write and execute code, 62% host sandboxes on Cloudflare, three times the closest competitor.
MIT Technology Review China reports that a16z crypto recently released a discussion featuring Stanford cryptographer Dan Boneh, Georgetown computer science professor Justin Thaler and others on the potential threat that rapid gains in AI math capability pose to modern cryptography. The report notes that OpenAI previously released hundreds of mathematics research manuscripts generated by internal models, but the list did not include cryptography results; Boneh said the public results did not show an actual break of mainstream cryptographic schemes and argued frontier labs should disclose whether their strongest models tested deployed systems and what the results were. The discussion also covered Ethereum Foundation researcher Justin Drake's "bunker mode" proposal, parallel deployment of post-quantum standards, and "crypto agility," while industry remains divided on when attacks might occur and how migration should proceed.
Origin now integrates more than 20 apps spanning deployments, CI, code review, security, monitoring and project management, with partners including Cloudflare, Vercel, Netlify, CircleCI, CodeRabbit, Greptile, Semgrep, Socket Security, Sentry and Linear, according to the Cursor changelog. The marketplace is also open to public submissions starting today, so any team building developer tools can publish an app for Origin. The submission flow asks developers to create an app in their namespace, build against the Origin API — authenticating with an app JWT and installation access tokens, receiving webhook events and writing back check runs, pull request comments and reviews — then make it public and submit a slug, suggested tags and a demo video for review before listing.
This article, reposted from the Agentic AI Foundation blog, explains WebMCP, a proposed standard that lets websites register structured tools through document.modelContext so AI agents can call site functions with defined input schemas. It shows both script-based registration and a declarative HTML form path, describes how WebMCP differs from MCP and MCP Apps, and stresses that tools should still use existing access controls and input validation. Experimental implementations exist in Chrome through a flag and origin trial, a Cloudflare developer preview, and Brave's Leo assistant, while a June preprint tested tool-surface poisoning in a polyfill environment.
An opinion piece in TechRadar Pro's Perspectives channel argues that as autonomous machines spread through cities, farms and factories, the data they generate — and the updates they need — may become a bigger constraint than the machines themselves. It cites FAA projections that the US commercial drone fleet could grow from about 966,000 aircraft at the end of 2024 to roughly 1.18 million by 2029, and estimates that 4.66 million industrial robots were working in 2024, with about 500,000 installed that year. The author says the fastest way to move large volumes of data off a machine in 2026 is often still a person carrying a hard drive, and cites Google DeepMind's Gemini Robotics 1.5 as evidence that skills can transfer zero-shot between physically different robots. The proposed fix is wireless optical communication, described as carrying fiber-like capacity of up to 20 Gbps over as far as 20 kilometers with clear line of sight, paired with an RF backup path because fog and rain can degrade optical links.
Talorys is an open-source, single-user personal AI assistant deployed with one command (npx create-talorys@latest) into the user's own Cloudflare account, with no server or account operated by the developers. It offers streaming chat, durable memories that can be viewed, edited and deleted, task/note/project management, and one-off or recurring reminders and routines. Chat runs on Workers AI's @cf/zai-org/glm-4.7-flash; when the free daily AI allocation is used up, chat pauses but tasks, notes, memories and reminders keep working.
CNBC commentator Jim Cramer said the surge in network traffic generated by AI agents is creating growth opportunities for Cloudflare, Akamai and Fastly, all of which have outperformed the S&P 500's roughly 12% gain this year (Fastly +159%, Cloudflare +78%, Akamai +22%). He said Cloudflare and Fastly look too expensive or too volatile to buy now, and called Akamai his only strong pick, citing a forward P/E below 16 and Anthropic's September 24 commitment to spend $11.6 billion with Akamai over seven years for additional compute capacity.
TypeSafe AI emerged from stealth on September 15 with Jev, a model that does not produce free-form text but instead returns estimated probabilities over a fixed answer set — yes/no, multiple choice, or a rating. It went viral: the announcement drew nearly 40 million views on Twitter, over 100,000 people joined its Discord within a week, some 2,000 GitHub projects adopted it, and signups opened on September 20 were paused two days later amid overwhelming demand. Because it is faster and cheaper and its output slots directly into code patterns like if statements, Jev has spawned similar models from the likes of Cloudflare, Amazon and OpenAI.
OpenRouter announced on X that Cloudflare's Clef (27B) and Clef Flash (9B) open-weight decision models are now available through OpenRouter. According to the post, the models accept text, JSON or images and return typed answers with probabilities instead of generated text. Pricing is listed at $0.24 per million input tokens for Clef and $0.09 per million for Clef Flash, with output free.
TypeSafe AI’s Jev is a decision model that maps inputs to predefined outputs such as yes/no answers, scores, or list selections instead of generating free-form text. CEO Diogo Almeida claims the model is used by about 25% of Fortune 500 companies and processed one trillion tokens per day roughly a week before the report, while separate reports say the startup is discussing a new funding round; these claims were not independently verified in the source.
What problem it solves
Cloudflare WARP does not provide a way to choose the edge node used by a tunnel, and an unsuitable node can prevent sites from opening even when WARP connects. Users need a way to discover endpoints that work from their network, identify their node and exit region, and produce a usable configuration.
Key features
Endpoint discovery — Scans Cloudflare WARP endpoint addresses …
What problem it solves
Exposing an origin through Cloudflare Tunnel can require inbound network access, which conflicts with keeping the origin closed. The client connects the Cloudflare network to the origin and also supports access to protected origins for Layer 4 traffic such as SSH and RDP.
Key features
Cloudflare Tunnel connectivity — Runs a tunneling daemon between the Cloudflare network and…
What problem it solves
Tools like Semrush or Ahrefs are described as expensive and bloated, which puts comprehensive SEO tooling out of reach for many people. OpenSEO aims to replace them with focused workflows, self-hosting options, and a pay-as-you-go model where you bring your own DataForSEO API key.
Key features
Core SEO workflows in one tool — Keyword research, rank tracking, competitor insig…
What problem it solves
Personal and home networks often require several separate tools for dynamic DNS, local service access, certificates, file services, and device automation. Lucky brings these common networking tasks into one interface and originated as a way to forward traffic from public IPv6 to private IPv4 devices on a router.
Key features
Domains and Service Access — Manage dynamic DNS, r…
What problem it solves
Leaked credentials can be difficult to identify and assess because they may be stored across repositories, chats, wikis, logs, filesystems, and other systems. Finding a credential pattern alone does not establish whether it is active, so the project combines discovery, classification, validation, and analysis.
Key features
Broad secret discovery — Scan sources including Git …
What problem it solves
Professionals and teams need control over their email and related productivity tools. The project provides custom-domain email, calendar, drive, booking, and built-in AI agents in a self-hosted application.
Key features
Custom-domain email — Create personal or shared mailboxes on a custom domain and send and receive mail through supported providers. Incoming mail can be stor…
What problem it solves
Security audits require systematic coverage, adversarial validation, and reliable source claims. This skill structures those activities into multiple phases and separates confirmed findings from unresolved or disproved candidates through machine-readable records.
Key features
Six-phase audit workflow — The audit progresses from reconnaissance through coverage-led hunting, ca…