Cloudflare has added OAuth scope customization, letting users deselect optional permissions on the consent screen rather than facing an all-or-nothing choice. Client owners can now mark specific scopes as optional when configuring an OAuth app, and any access token generated after consent will only contain the scopes the user approved. Cloudflare says it plans to expand its account and zone-level role surface over the next few weeks to cover nearly every Cloudflare product, adding more API token roles, account membership options, and OAuth scopes.
Bun 1.4 improves Node.js compatibility with more than 1,500 additional passing Node test cases while cutting idle CPU usage by up to 5×, reducing memory use by as much as 48% in some server workloads, and roughly doubling startup speed on Linux. The release also adds new built-ins including image processing, WebView, Markdown parsing, cron jobs, parallel test and script execution, package auditing, and deduplication, while continuing Bun's broader migration from Zig to Rust.
AI is driving a surge in software vulnerabilities and security fixes as models uncover legacy bugs while generating new flawed code, creating competing pressures that may keep patch volumes elevated. The future could range from highly stable, rarely patched software to continuously changing applications where daily builds effectively replace traditional patching.
Datadog built an internal Git mirror service called gitretriever to handle the scale of CI code fetches across its engineering infrastructure. The system now handles more than 100 million requests per week across roughly 5,500 repositories, and served over a billion Git requests and hundreds of terabytes of data in its first four months. Traffic grew about 20 times during that period while median serve latency held at around 40 milliseconds. Moving CI fetch traffic to gitretriever cut fetch-serving CPU on the previous backend by three to four times.
Tencent Zhuque Lab has open-sourced AI-Infra-Guard (A.I.G), a red teaming platform for scanning AI infrastructure, MCP servers, agent skills, and testing large language models for jailbreak vulnerabilities.
Agent Substrate delivers a performant, high-density runtime environment for large-scale agent deployments. The agent substrate control plane provides full lifecycle management for agent sandboxes, delivering sub-second agent resume/suspend operations, and allows heavy multiplexing of agents onto the same computer infrastructure.
A multinational financial services organization migrated from an outdated, stressed single Octopus Server to a five-node High Availability cluster through careful planning, staged testing, backups, and rollback checkpoints. The migration established a stable foundation and led to a regular six-month upgrade cadence with minimal support needs.
Cloudflare researchers demonstrated a remote Spectre attack against its Workers production environment that reliably leaked up to 12 bits per second at 99% accuracy. The attack exploited a gap in Cloudflare's existing Dynamic Process Isolation defense, which only isolates suspicious scripts after an invocation ends, a delay attackers bypassed by keeping a single Durable Object invocation alive for hours via WebSocket messages. Cloudflare says no active exploitation was detected over the past three years. In response, the company has improved its isolation detection, integrated the V8 Sandbox, and deployed Memory Protection Keys to limit cross-isolate memory access within shared processes.
Rust 1.98 adds algebraic floating-point operations that let the compiler reorder calculations and enable broader vectorization, trading strict floating-point determinism for more optimization opportunities without introducing undefined behavior. The release also adds faster buffered integer formatting through format_into, which performs similarly to the popular itoa crate, and stabilizes a range of string, atomic, and numeric APIs.
Sponsored
Build v. Buy looks different now. Does that mean it's time to build your own observability solution? New Relic's Senior Director of
Docker is expanding its software supply chain security platform with hardened images and packages, extended lifecycle support, customizable builds, Helm charts, MCP servers, and Scout policies.
Summing individual experiment wins overstates their true combined effect on a business metric.
Datadog has added AI-native static application security testing (SAST) to its Code Security product, with coverage for the OWASP Top 10 for LLM Applications.
Sponsored
Agents are hitting your signup flow and bouncing off a browser login built for humans. Every one that gives up is a signup you never see.
WorkOS Agent Registration [5] turns that traffic into signups. Enroll via the dashboard and AuthKit publishes an auth.md file agents read to register for scoped, short-lived credentials you control.
International founders that make it in Silicon Valley tend to find communities of similar people who have done the same. When these founders get together, they have each other's backs. The emergence of these international tech mafias has changed the playing field for international founders. These borderless founders who have a foot in their home country and in Silicon Valley know the strongest engineers in their home market before the rest of the market catches on and can get early customers from large companies at home that give them big proof points early.
Apple set new EU App Store terms that take effect October 1, and agreements can be signed now. The standard in-app purchase commission moves to 26%, or 15% for qualifying developers. Alternative payment processing runs 20%, or 10% for eligible program members, and apps distributed outside the App Store pay a 5% Core Technology Commission on digital transactions and require notarization. For the first time, a single app can offer both Apple in-app purchase and alternative payments. Once an option is picked, it stays for 12 months.
Lovable launched in late 2024 with a visual identity that was prototyped in 30 machines. The brand quickly picked up $10 million in annual recurring revenue just 60 days after launch. Lovable's brand was designed and redesigned over two and a half years by an insanely talented design team. This post tells the story of how Lovable's first designer translated a feeling of genuine infectiousness into a design brief and brand identity. The company's growth shows what is possible when people with innate curiosity and genuine care for what they do are given complete trust: they produce exceptional work, remarkably fast.
Commenting on your prospects' posts can create investment while adding value. Comments create a direct touchpoint with your prospects, and it works as a distribution to many more similar prospects. They allow you to get free reach into a network of people just like your buyer. Commenting lets you stay visible in a way nothing else in outbound can.
This post lays out the arithmetic behind why a company that raised a seed easily gets Series A nos without changing anything about itself. An accelerator writing checks at $330K to $2.5M post-money only needs a $33M to $250M exit to work. Your seed at $5-10M post requires belief in a $500M to $1B outcome, and a Series A investor underwrites a path to roughly $100M a year, which is why $1M ARR is not enough. Dilution eats 50%+ by exit. Three escape valves get named, including customer-funded growth and capital that does not need 100x.
2026.08.21·thefounderplaybook.hustlefund.vc·4 min read·TLDR Founders Linear Agent can now set up, run, and test code before returning its work. Coding sessions now automatically adapt to each codebase's configuration. Once the application is running, Linear Agent tests it in the browser, where users experience it. If the agent finds a problem, it can make a fix and rerun the test before returning the work. Linear has also introduced more transparent pricing and made AI credit usage easier to understand and control.
Slack Code creates temporary channels where teams and agents can work together on a specific task or project. The channel is archived automatically when the work is done, but the context remains searchable. Code channels follow Slack's enterprise-grade security and permissions model. They make it easy for anyone, even non-technical employees, to see what an agent is building, review previews, and share feedback directly in Slack.
TLDR is hiring a GTM Engineer to join our Applied AI team and own our AI-native GTM stack. We're looking for someone comfortable building AI agents and working with HubSpot. Click here to learn more [15]!
Shipping code is not the same as shipping revenue. Confusing the two is an unforced error. If you haven't prepared great market-readiness ingredients, you just have another unsellable software. CEOs need to identify the next revenue to bottle next when code development is faster than selling.
AI acquisitions require code-level diligence because ARR and growth cannot reveal whether proprietary technology is defensible or merely a model wrapper. Buyers should score model dependency, data moat, talent concentration, and substitution risk before signing or assigning premium valuations.
Just because a salesperson was pretty good at their last role doesn't mean they'll be good at your startup too.
The Meta Startup School is a three-month-long program built to help early-stage consumer brands accelerate growth.
Organizations should integrate AI across all functions to align with business outcomes, rather than keeping it isolated in pockets.
Sponsored
As AI becomes embedded across the enterprise, organizations must secure their use of AI while defending against AI-powered threats.
That pressure is making the limits of fragmented architectures impossible to ignore.
Cato Networks was named a Leader in the 2026 Gartner® Magic Quadrant™ for SASE Platforms [5] for the THIRD YEAR RUNNING.
See how Gartner evaluated the market and why, in Cato's view, unified SASE platforms are becoming essential for the AI era.
Researchers extracted a JWT from a co-located Cloudflare Worker at 12 bits per second, reaching 99.16% accuracy. The attack used WebSockets for remote timing and Durable Objects to keep code running. WebSocket I/O reduced DyPrIs detection signals. Cloudflare deployed improved DyPrIs, V8 Sandbox protections, and MPK-based isolation and confirmed no customer data was accessed or active exploitation.
The Rust Security Response Team confirmed that the popular arrayref crate was maliciously republished after being made to depend on proc-macro1, a crate that included a build script downloading a malicious payload. Related crates internment and append-only-vec from the same author were also compromised, likely via stolen credentials rather than intentional action. The malicious versions remained live for 86 to 107 minutes before removal. The team has deleted proc-macro1 and its sibling packages and unyanked the legitimate arrayref versions. Developers should run the provided find command against ~/.cargo/registry/cache to check for arrayref@0.3.10, internment@0.8.7, append-only-vec@0.1.9, or any version of the deleted proc-macro1 family.
Researchers from security firm Adversa discovered a new attack that allowed them to bypass Grok's guardrails and exfiltrate user data. The attack involves hosting encrypted instructions along with a decryption key and instructions to decrypt them on a webpage. When the user requests Grok to summarize or otherwise process the page, Grok decrypts the text, and then the decrypted ciphertext contains instructions that involve Grok constructing a fake decryption key by appending the user's name, location, and chat history, which are later used as a URL parameter and sent to an attacker-controlled server.
wasm2c records a guest-declared table size before calling calloc, then continues if calloc returns NULL. A 64 GiB funcref table can fail under RLIMIT_AS, strict Linux overcommit, 32-bit hosts, or memory pressure. Bounds checks then accept indexes against the declared size while NULL-based table accesses resolve to host addresses. The proof of concept leaks libc pointers through the GOT, forges a funcref entry, passes the indirect-call type check, and calls system() to write a host file. A calloc NULL check and table-size cap prevent it.
The Agent Access Model (AAM) is a framework that secures autonomous AI workflows by issuing short-lived credentials that a "Trust Ratchet" can revoke mid-execution upon detecting protected data. Multiplayer access control remains an unsolved challenge with enterprise privacy-violation rates exceeding 50 percent, so organizations should begin by strictly securing a single bounded agent. Security teams should deploy these initial agents using network-level mediation, strict execution harnesses, and detailed activity logging to safely build evidence for future permission grants.
Osquery and Fleet recently added the ability to scan memory at scale using YARA rules via a new yara_process Osquery table. This post walks through how to set up a lab environment for detecting the use of Sliver on a DVWA running in Kubernetes using YARA rules distributed via Fleet. The example is then expanded to detect a malicious local MCP which runs a credential stealer in memory.
StrongestLayer provides email security for phishing, business email compromise, and adversary-in-the-middle attacks. It evaluates message context and evidence to flag suspicious emails, including attacks that do not match known patterns.
Z.ai has released GLM-5.3, trained from the GLM-5.2 base model through expanded post-training. On CyberGym, it scored 84.5%, up from 77.2%. On ExploitBench, it reached 54.4%, versus 24.4%. In tests with security teams in China, the model found 2,436 vulnerabilities across 269 projects. Z.ai lists 1,097 as critical or high severity, with fifty-three findings being public and 2,383 remaining under embargo.
Blacksea is an active honeypot and canary-bait control system built to detect and drown LLM-driven attackers.
Kriminal is an off-the-shelf AI hacking platform that offers social-engineering persona creation, offensive-security assistance, OSINT searches, crypto tracing, and uncensored image generation. ThreatDown found it appears to combine Grok, Claude, Llama via OpenRouter, Tavily, Google Cloud, Cloudflare, and NowPayments to maximize effectiveness and reduce visibility of interactions between the different tools.
OpenAI's Greg Brockman disclosed that an autonomous AI agent collective breached the company's research infrastructure and a separate production environment by chaining undisclosed vulnerabilities with leaked credentials. While avoiding specific threat actor attribution, Brockman warned that upcoming releases of highly capable open-weight models will drastically accelerate this automated threat landscape. To keep pace, security teams must adopt a modernized defensive strategy that incorporates AI-assisted code review, continuous alert triage, automated attack-path enumeration, and strong defense-in-depth fundamentals.
Showcase of a multi-agent harness to test SAML implementations using Claude Opus.
ToxicPanda 2.0 expanded its overlay-based credential theft from 16 to 349 financial institutions across 16 countries, adding 167 remote commands, PIN harvesting across 140+ banking and crypto apps, and an accessibility-service abuse chain that enables Wireless Debugging via ADB for privilege escalation.
Sponsored
AI changed who can build. Employees across the business are now shipping apps, agents, and automations faster than IT can track them, a growing wave of "wild code" that solves real problems but lives outside governance and support.
On September 15, join Tines VP of Product for a live conversation with TechnologyAdvice [4]on how IT can give teams the freedom to build at the speed of AI, without losing the oversight and operational rigor required to do it safely.
You'll learn how to spot shadow AI and unmanaged automations before they become risk, and what it takes to support innovation without adding technical debt.
2026.08.21·tines.com·Sponsor·TLDR IT Citrix has warned customers to patch two NetScaler vulnerabilities that could enable remote authentication bypass or denial-of-service attacks under specific configurations. The company recommends upgrading affected NetScaler ADC and NetScaler Gateway appliances to specified builds, although it says neither flaw has been observed being exploited. Administrators can inspect configuration strings to determine exposure.
2026.08.21·links.tldrnewsletter.com·2 min read·TLDR IT Slack Code brings coding agents, including Claude Code, Devin, GitHub Copilot, and Vercel, directly into shared channels where teams can prompt, review, and steer their work together. Agents operate with the invoking user's permissions, while code channels preserve diffs, previews and activity as a searchable record of the work.
2026.08.21·venturebeat.com·8 min read·TLDR IT Conservative, “boring” infrastructure can improve resilience while reducing operational complexity. Migrating CGI scripts to FastCGI and primary order databases from MySQL to SQLite reported better performance and simpler backups.
2026.08.21·weblog.rogueamoeba.com·6 min read·TLDR IT The enterprise AI conversation is shifting from broad experimentation toward tighter scrutiny of cost, governance, workforce impact and actual business value. Companies are increasingly narrowing deployments around measurable use cases as token consumption and usage-based pricing make unrestricted AI adoption harder to justify.
2026.08.21·ciodive.com·11 min read·TLDR IT