# want Published articles for want. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## AI workflow demo - create a skill with evals in antigravity DevFeed: [AI workflow demo - create a skill with evals in antigravity]() Original publisher: [Read original article]() Published: 2026-11-01T17:01:30Z Content type: tutorial Language: en Sources: [Maiatoday]() Topics: [Test automation](), [Agent Skills](), [google-antigravity]() Tags: [ai](), [antigravity](), [evals](), [prompt](), [repo](), [resources](), [skill-creator](), [skills](), [summary](), [test](), [testing](), [want](), [work](), [workflow]() ### AI overview The author recommends pairing important skills with evals so their behavior can be checked, repeated, and revisited. Evals record triggering prompts, expected outcomes, and optional fixtures or code snippets. In the Antigravity demo, Anthropic's `skill-creator` builds a repository-history catch-up skill and two eval cases; the author reviews and edits the evals, then runs them with and without the skill using the currently loaded model. The reported result suggests the skill helps in that scenario. Evals are also presented as a way to check whether newer models can do the task without the skill. For reusable skills, the author found Antigravity's `~/.gemini/config/skills` location unavailable to other agents, so chose `~/.agents/skills` with symlinks to agent-specific locations. The Anthropic version of `skill-creator` stores results in a sibling `-workspace` folder, which the author says Claude Code ignores and some agents may error on; they changed it to store results beside the evals. ### Source excerpt What When I create an important skill for me or my team it is useful to know if it works. To do this I also create an eval that contains what prompt is used to trigger the skill and what the outcome should be. The eval can also contain fixtures and sample code snippets to work on. This workflow isn't specific to antigravity but I tested it there. Why Sharing If I am sharing skills with other people I want to make sure they work as expected. Initially I would test them by hand a few time but I would have to rely on my memory to know if the work. Also this manual testing takes a long time. With evals I can automate these steps. Model changes Models become more powerful and sometimes the skills become obsolete because newer models know how to do the tasks. With an eval I can see if the model can perform better than the skill. If it does then the skill can be deleted and the context freed up. Future me If I come back to a skill after some time, sometimes I can't remember how it worked. If I read the eval I can see what I expected it to do Resources A useful open source skill that you can use to create skills and evals and run evals is skill-creator from anthropic. A good tutorial explaining how to create a skill including the evals. General skills documentaion Demo Creating a skill and an eval Use the skill-creator to create a new skill. The detail of what the skill does doesn't matter. How you ask for it and checking that the evals are created is the most important part of this workflow. For our example I am creating a skill to pull the history on a repo after I come back from some time away touching grass. The prompt is this use skill creator to create a user level skill that will pull the history of a git repo and give me a summary of what happened while I was away. The time period could be anything but typically 6 weeks. Summarise by type of tasks, features, chores and show who did the work. Also group related tasks. I want to use this to get a scanable overview of ## Here's How You Can Use VPNs for FREE on Linux DevFeed: [Here's How You Can Use VPNs for FREE on Linux]() Original publisher: [Read original article]() Author: Abhishek Prakash Published: 2026-11-01T15:07:36Z Content type: article Language: en Sources: [It's FOSS]() Topics: [Linux](), [Firefox](), [Linux Mint]() Tags: [2](), [2013](), [2016](), [browser](), [but](), [can](), [don-t](), [either](), [free](), [here](), [here-s](), [linux](), [list](), [need](), [options](), [plan](), [system](), [vpn](), [want](), [wide](), [work]() ### AI overview Free Linux VPNs differ most in coverage and limits: browser options protect only browser traffic, while native clients cover the whole system. Vivaldi's Proton integration and Opera are unlimited for browsing; Firefox offers 50 GB monthly where available. Opera needs no account but lets users select only a broad region, and acts as a browser proxy. For system-wide use, Proton VPN has no data cap, but assigns the free server, supports one device, and excludes streaming and P2P. Windscribe allows users to choose among 10 countries and use unlimited devices, with 10 GB monthly after email confirmation (2 GB without). hide.me requires no sign-up and has unlimited but heavily throttled data; AdGuard VPN offers 3 GB monthly at up to 20 Mbps. RiseupVPN is donation-funded with no account, while Lantern provides about 250 MB daily for most users. The article recommends paid service for regular use or streaming, and notes free plans generally have limited locations or speeds. ### Source excerpt Need a VPN once in a while but don't want to pay for a yearly plan? Here are free VPN options that work on Linux, either in the browser or system-wide. ## Five data tricks you don't want this Halloween DevFeed: [Five data tricks you don't want this Halloween]() Original publisher: [Read original article]() Author: Francisca Segovia-Garcia Published: 2026-10-30T13:59:10Z Content type: article Language: en Sources: [Veeam Software Official Blog]() Topics: [Disaster Recovery](), [enterprise data protection](), [Security](), [ai security](), [trust]() Tags: [ai](), [avoid](), [business](), [can](), [complexity](), [course](), [data](), [data-governance](), [data-resilience](), [data-security](), [days](), [deal](), [doesn-t](), [don-t](), [enterprise](), [enterprise-data](), [first](), [full](), [fun](), [halloween](), [help](), [ransomware](), [this](), [tricks](), [want]() ### AI overview The article argues that backups alone do not ensure reliable recovery, secure data, or trustworthy AI. It identifies five risks: backups may not restore the right data cleanly and in time; protected copies may still be exposed to ransomware; organizations may lack visibility into data sensitivity, permissions, and appropriate uses; fragmented tools and dashboards can increase operational complexity; and unreliable data can undermine AI. It recommends precise recovery of clean data, combining security with protection, connecting context across data and access, using intelligence and automation to prioritize action, and governing and protecting data used by AI. The article cites surveys on recovery, ransomware, AI governance, incident impacts, and AI readiness to illustrate these concerns, and promotes Veeam's platform approach. ### Source excerpt The deal is simple: Give us a treat, or there might be a trick. Of course, these days, we're mostly there for the treat: A chocolate bar, a handful of candy or, if you picked the right house, a full-size one. The possibility of a trick adds to the fun, even though most of us expect a treat. If only enterprise data worked the same way. A recovery that doesn't recover. Sensitive data where you didn't expect it. Security tools that add complexity. AI built on data you can't confidently trust. Nobody wants surprises in these areas. So, this Halloween, let's knock on five doors to uncover common data tricks and the treats that can help you avoid them. The post Five data tricks you don't want this Halloween appeared first on Veeam Software Official Blog. ## Part 3 - Cybersecurity journeys: The work behind the work DevFeed: [Part 3 - Cybersecurity journeys: The work behind the work]() Original publisher: [Read original article]() Author: Shannon Brazil Published: 2026-10-27T17:00:07Z Content type: article Language: en Sources: [AWS Security Blog]() Topics: [Cybersecurity](), [Artificial Intelligence](), [Incident response](), [Security Operations Center](), [ctf]() Tags: [17](), [2](), [3](), [about](), [admin](), [clocks](), [conversations](), [cybersecurity](), [i](), [important](), [just](), [make](), [quiet](), [reading](), [security-identity-compliance](), [show](), [television](), [that-s](), [this](), [thought-leadership](), [want](), [work]() ### AI overview Cybersecurity work depends on patient practice, communication and judgment as much as technical skill. Application security engineer Brandon Caldwell reviews code while products are being built, then explains flaws empathetically to engineers; he values clear writing and looks for practical evidence such as capture-the-flag results and public GitHub projects when hiring. Pen tester Kurt Heath recommends building foundations in systems, networking, development and security before pursuing offensive work. He uses AI to help with unfamiliar code and large information sets, but cautions against letting it act during engagements because it may cause damage or exceed scope. Application security engineer Seth Jeacopello describes the attention management required to review large codebases; his suggested exercise is to attack a deliberately vulnerable machine and inspect its logs. The article also points readers to TryHackMe, Hack The Box, OWASP resources, AWS Skill Builder and AWS Builder Labs for practice and learning. ### Source excerpt The most important work I've done in cybersecurity would make for terrible television. There were no countdown clocks and no frantic typing, just a lot of careful reading, a lot of hard conversations, and the occasional quiet moment where everything finally clicked. That's the part I want to show you this week, because the three [...] ## Silicon Valley parents want the tech they help build out of their children's schools DevFeed: [Silicon Valley parents want the tech they help build out of their children's schools]() Original publisher: [Read original article]() Author: Robin Buller in San Francisco Published: 2026-10-27T12:00:38Z Content type: news Language: en Sources: [The Guardian | Technology]() Topics: [Learning]() Tags: [2015](), [area](), [build](), [but](), [california](), [california-silicon-valley-us-education-education-children-children-s-tech-society-technology-us-news](), [children](), [children-s-tech](), [education](), [engineer](), [experience](), [facebook](), [gil](), [github](), [help](), [moving](), [netflix](), [one](), [private](), [silicon](), [silicon-valley](), [society](), [tech](), [technology](), [us-education](), [us-news](), [want](), [west-coast]() ### AI overview Tech industry parents in affluent California communities are pressing schools to reduce classroom screen use, citing concerns about learning, student-teacher relationships and young children's exposure. San Mateo schools stopped sending Chromebooks home with students in kindergarten through second grade and banned high schoolers' cellphones; Mountain View is moving away from i-Ready, while Palo Alto prohibited smart glasses. Parents also want districts to vet educational products and require evidence of learning outcomes. Private schools such as San Francisco Waldorf can change policies more quickly and devote more resources to teachers, but its tuition is high. Educators warn that lower-income districts may rely more heavily on screens when budgets are tight, while families leaving public schools for private options can take funding with them. The excerpt describes local efforts and concerns, not their long-term effects. ### Source excerpt Both private and public schools in tech regions are limiting the use of tech in classrooms. But will low-income districts be able to follow suit? German Gil spends most of his waking hours looking at screens. Since moving to the San Francisco Bay Area in 2015, the software engineer has bounced between startups and giants including Netflix, Facebook and GitHub. But when it came to the education of his daughter, Victoria, he wanted her to have a different experience, one in which his "kid could be a kid". Continue reading... ## How Much of Your Software Factory Can Run in the Dark? DevFeed: [How Much of Your Software Factory Can Run in the Dark?]() Original publisher: [Read original article]() Author: stack72 Published: 2026-10-22T20:38:34Z Content type: opinion Language: en Sources: [Swamp Club Feed]() Topics: [engineering-culture](), [implementation](), [human review](), [Test automation](), [human computer interaction](), [Developer experience](), [Integration testing]() Tags: [about](), [account](), [actions](), [added](), [agent](), [build](), [building-with-ai](), [builds](), [can](), [comes](), [factory](), [giving](), [let](), [responsibility](), [run](), [software](), [software-factory](), [systems](), [want](), [with](), [work]() ### AI overview The article argues that operational software can run unattended when its job, authority, response boundaries and escalation path are defined, while product development often needs human involvement to shape behavior and revise assumptions. Agents can investigate and implement features, but seemingly clear tickets leave consequential choices--for example, whether workflow retries restart from the beginning or resume after failure. Discussing such choices during development helps avoid embedding unwanted behavior in code, tests and documentation. The author also recommends using agents to explore and try alternatives, not simply ship the first working version. Both kinds of work need independent verification: agent-written tests can confirm a mistaken interpretation, and automated checks cannot determine whether the intended behavior was chosen. The boundary shifts as operational incidents expose new problems or product decisions become settled. Repeated questions should be captured in context or checks; human involvement is most useful where decisions still need discussion. ### Source excerpt Our lights-on software factory builds systems we want to operate with the lights off. We work through the responsibility we're giving those systems, build it into the operation and let them run. When something new comes back, we have more work to do in the software factory. ## Personal Data for Millions of California Students Stolen in Massive Breach DevFeed: [Personal Data for Millions of California Students Stolen in Massive Breach]() Original publisher: [Read original article]() Author: EditorDavid Published: 2026-10-17T17:00:00Z Content type: news Language: en Sources: [Slashdot]() Topics: [Security](), [Access Control](), [Credential theft](), [sensitive data](), [Threat monitoring]() Tags: [breach](), [california](), [data-breach](), [education](), [information-security](), [personal-data](), [want]() ### AI overview A breach involving Educational Testing Service exposed personal information, including names, birth dates, and addresses, belonging to millions of current and former California public-school students. A compromised login credential was used to copy student files between May 22 and September 18, 2026. A cybersecurity expert raised concerns about authentication, access restrictions, and monitoring; ETS said it has strengthened safeguards and added monitoring and third-party security reviews. ### Source excerpt Millions of current and former students at California's public school have had their data stolen -- including names, date of birth, and addresses -- stolen in a breach, reports the Los Angeles Times. The breach affects students "who took the state's mandated standardized tests that annually assess their academic proficiency, education officials announced Friday." The compromised information affects many more than the state's current 5.7 million public-school students as the breach dates back to 2023 enrollments. Students who left California but did not confirm their transfer to another school also were affected, dating back to 2006. The data breach involved [nonprofit] Educational Testing Service or ETS, which provides extensive services to California [including test participation and assessment data]... Affected former students, including those who are now adults, are eligible for 24 months of free identity-protection services and identity-restoration help if there is identity theft. Officials said that parents and guardians may want to consider requesting a free credit freeze for their children who are under 16 to help prevent unauthorized accounts from being opened in the names of their children. The breach involved a compromised login credential, according to a web page about the incident from California's Department of Education. The compromised credential was then used to copy files with student information at "various" times between May 22 and September 18, 2026, which drew concern from a cybersecurity expert interviewed by the Los Angeles Times: "A stolen login credential should not become a months-long passport to millions of students' personal records," said Anton Dahbura, executive director of the Johns Hopkins University Information Security Institute. "If that is what happened here, the critical question is why authentication safeguards, access restrictions, and monitoring systems failed to detect and stop the unauthorized activity." He added: "Incidents ## Elecrow CrowPanel Advanced HMI Display Review: A Tinkerer's Touchscreen DevFeed: [Elecrow CrowPanel Advanced HMI Display Review: A Tinkerer's Touchscreen]() Original publisher: [Read original article]() Author: Abhishek Prakash Published: 2026-09-19T14:25:32Z Content type: article Language: en Sources: [It's FOSS]() Topics: [touchscreen](), [Nesso N1](), [ipad]() Tags: [2](), [accessible](), [advanced](), [ai](), [board](), [control](), [control-panel](), [display](), [doing](), [don-t](), [elecrow](), [gadgets](), [hmi](), [homelab](), [makers](), [mind](), [panel](), [review](), [reviews](), [touchscreen](), [want](), [work]() ### AI overview The Elecrow CrowPanel Advanced 10.1-inch is a maker-oriented ESP32-P4 touchscreen board that can serve as a Home Assistant control panel, but requires custom firmware and setup. The reviewer flashed ESPHome, which communicates through Home Assistant's native API and worked with a container-based installation; a test button toggled a geyser plug and stayed synchronized with changes made in the app. Initial Wi-Fi setup required correcting reversed SDIO data-line order on the reviewer's V1.1 unit, while the stock demo firmware did not retain Wi-Fi credentials after reboot. The 1024x600 IPS screen had reliable touch response and adequate brightness for wall controls, though its image quality was compared to an entry-level tablet. Speakers, microphone, camera connection, battery connector, and exposed I2C, UART, and GPIO offer room for projects; battery life and on-device AI were not tested. The reviewer recommends it for people comfortable building and configuring projects, not those seeking a ready-to-mount panel. ### Source excerpt A big touchscreen board for makers who want a physical control panel for their homelab, if you don't mind doing the work yourself. ## strace-ui, Bonsai\_term, and the TUI renaissance DevFeed: [strace-ui, Bonsai\_term, and the TUI renaissance]() Original publisher: [Read original article]() Author: James Somers Published: 2026-05-26T00:00:00Z Content type: article Language: en Sources: [Jane Street]() Topics: [Text-based user interface](), [debugging](), [debug tools](), [Bonsai]() Tags: [but](), [can](), [debugging](), [exploring](), [filter](), [flag](), [follow](), [found](), [it](), [it-s](), [make](), [this](), [threads](), [tool](), [trace](), [ui](), [useful](), [want](), [with](), [work]() ### AI overview strace-ui makes strace traces easier to scan and explore interactively. It assigns short labels to processes and threads, formats structs, and displays buffers as hexdumps. Users can hide syscalls, jump between calls using the same file descriptor, filter by FD, PID, or excluded PID, and open a syscall's man page. It attempts to track FD reuse and follow descriptors across forks, but the excerpt says this works poorly when attached to a process that already has open FDs. DNS resolution can display hostnames alongside ports. The excerpt then introduces Bonsai, an OCaml framework for composing interactive UIs from purely functional state machines, as part of the context for strace-ui's development; it cuts off before explaining further. ### Source excerpt We've always found strace useful but somewhat hard to work with. Its output is often inscrutable, it's hard to follow subprocesses or threads, and if you want to filter syscalls you have to rerun the trace with a flag for each one. What you want in debugging is a tool for exploring, refining, etc., but strace can make this difficult. ## Stream of Consciousness Driven Development DevFeed: [Stream of Consciousness Driven Development]() Original publisher: [Read original article]() Author: Hillel Wayne Published: 2026-02-18T16:33:08Z Content type: article Language: en Sources: [Newsletter feed for Hillel Wayne's Newsletter]() Topics: [software-development](), [Spec Driven Development](), [pair\_programming]() Tags: [chain-of-thought](), [client](), [computational-complexity](), [development](), [experience](), [explain](), [markdown](), [pairing](), [partner](), [question](), [quicksort](), [recursion](), [solution](), [sort](), [spec](), [want]() ### AI overview When a spec change is conceptually complex, write out the reasoning together before editing the spec. The author created a Markdown file covering the problem, failed attempts, proposed solution, how it works, expected changes, and what it would not address. Questions and dead ends were added to the document as the discussion unfolded. They began changing the spec only after the client understood the reasoning and agreed with the approach. The author says this helped prevent a partner from deferring to the perceived expert and leaving behind a spec they could not maintain. The approach had been tried once and seemed promising, but requires substantial typing. ### Source excerpt This is something I just tried out last week but it seems to have enough potential to be worth showing unpolished. I was pairing with a client on writing a spec. I saw a problem with the spec, a convoluted way of fixing the spec. Instead of trying to verbally explain it, I started by creating a new markdown file: NameOfProblem.md Then I started typing. First the problem summary, then a detailed description, then the solution and why it worked. When my partner asked questions, I incorporated his question and our discussion of it into the flow. If we hit a dead end with the solution, we marked it out as a dead end. Eventually the file looked something like this: Current state of spec Problems caused by this Elaboration of problems What we tried that didn't work Proposed Solution Theory behind proposed solution How the solution works Expected changes Other problems this helps solve Problems this does *not* help with Only once this was done, my partner fully understood the chain of thought, and we agreed it represented the right approach, did we start making changes to the spec. How is this better than just making the change? The change was conceptually complex. A rough analogy: imagine pairing with a beginner who wrote an insertion sort, and you want to replace it with quicksort. You need to explain why the insertion sort is too slow, why the quicksort isn't slow, and how quicksort actually correctly sorts a list. This could involve tangents into computational complexity, big-o notation, recursion, etc. These are all concepts you have internalized, so the change is simple to you, but the solution uses concepts the beginner does not know. So it's conceptually complex to them. I wasn't pairing with a beginning programmer or even a beginning specifier. This was a client who could confidently write complex specs on their own. But they don't work on specifications full time like I do. Any time there's a relative gap in experience in a pair, there's solutions that are concep ## How Not to Fight with Product Managers - as a Developer DevFeed: [How Not to Fight with Product Managers - as a Developer]() Original publisher: [Read original article]() Author: Anis Khan Site Reliability Engineering is my role Cost optimization is my goal GitHub profile Linkedin profile Published: 2026-02-02T00:00:00Z Content type: article Language: en Sources: [Trivago]() Topics: [Product Management](), [Maintainability](), [CI/CD]() Tags: [2](), [3](), [about](), [adopting](), [also](), [comes](), [developer](), [developers](), [engineering-culture](), [fast](), [feature](), [fight](), [giving](), [it-s](), [little](), [manager](), [product](), [product-managers](), [release](), [site-reliability-engineering](), [want](), [with]() ### AI overview Service Level Objectives (SLOs) give developers and product managers a shared, user-focused way to weigh feature delivery against reliability. Teams define measurable targets such as availability or latency, then use the remaining error budget to guide release decisions: with a 99.9% availability target over 30 days, the text calculates about 43 minutes of permitted downtime. Trivago describes tracking availability, latency, and business indicators, with alerts and dashboards; it pauses feature work when services are often unavailable and allows more experimentation when they remain stable. The article recommends starting with observed service performance, setting a realistic target, and revising it iteratively. It cautions that 100% reliability can require effort that does not pay off, so teams should agree on what is good enough for users. ### Source excerpt It's a scenario developers relate to a little too well. The product manager always comes with more and more feature requests. They also want to release fast by giving a tight deadline. While you... ## Our response to the January 2025 Kudelski Security vulnerability disclosure: action and continuous improvement DevFeed: [Our response to the January 2025 Kudelski Security vulnerability disclosure: action and continuous improvement]() Original publisher: [Read original article]() Author: Harjot Gill Published: 2025-08-19T00:00:00Z Content type: article Language: en Sources: [CodeRabbit]() Topics: [Incident response](), [Security](), [code security](), [incident]() Tags: [2](), [2025](), [3](), [added](), [ceo](), [continuous](), [continuous-improvement](), [customer](), [customer-data](), [data](), [disclosure](), [english](), [i](), [improvement](), [quickly](), [reports](), [security](), [share](), [vulnerability](), [vulnerability-disclosure](), [want]() ### AI overview On January 24, 2025, Kudelski Security disclosed that CodeRabbit's RuboCop service was running outside its secure sandbox. The company says it disabled RuboCop within an hour, rotated relevant credentials and secrets within three hours, and moved it into the sandbox with a production fix within 12 hours. Its investigation found no evidence of customer data access or malicious activity, and RuboCop was the only service it found outside the sandbox. Follow-up measures include automated sandbox checks, explicit sandbox verification in deployment gates, a broader service audit, and recurring checks for engineers' required security training. The company also reports a VDP average first response under 24 hours and resolution within seven days. ### Source excerpt No customer data was accessed and the vulnerability was quickly remediated within hours of disclosure As the CEO, I want to address recent reports of a security vulnerability discovered in January 2025 by Kudelski Security researchers and share our immediate response, the steps we've taken since, and our ongoing commitment to security. ## Array 1.42.0: Get beta features via our roadmap! DevFeed: [Array 1.42.0: Get beta features via our roadmap!]() Original publisher: [Read original article]() Published: 2022-11-30T00:00:00Z Content type: release Language: en Sources: [PostHog]() Topics: [roadmaps](), [feature flags]() Tags: [about](), [added](), [agents](), [ai](), [ai-observability](), [all](), [also](), [analytics](), [app](), [application](), [array](), [beta](), [features](), [know](), [newsletter](), [roadmap](), [self-hosted](), [send](), [up](), [want]() ### AI overview PostHog 1.42.0 adds area charts for insights, normalizes URL and pathname breakdowns, and makes it easier to share session recordings at a specific time. It also launches a public roadmap where users can vote, subscribe for updates, and give feedback. Self-hosted users must run async migrations 0005-0007 before upgrading; if those migrations are unavailable on their current version, the excerpt recommends upgrading to 1.41.4 first, running them, then moving to 1.42.0. Other listed changes include recording playback and dashboard improvements and fixes for experiment metrics and group pages. ### Source excerpt Want to know more about what we're up to? Subscribe to our new newsletter , which we send once every two weeks! 1.42.0 update guide for self-hosted... ## Array 1.32.0 DevFeed: [Array 1.32.0]() Original publisher: [Read original article]() Published: 2022-01-31T00:00:00Z Content type: release Language: en Sources: [PostHog]() Topics: [User interface design](), [product-experimentation]() Tags: [2](), [2015](), [3](), [about](), [actions](), [actor](), [addition](), [agents](), [ai](), [ai-observability](), [all](), [also](), [array](), [easier](), [experimentation](), [funnels](), [groups](), [launch](), [vertical](), [want]() ### AI overview PostHog 1.32.0 redesigns Persons & Groups pages, shortens shared insight URLs, improves event and property discovery when building insights, and adds vertical funnel breakdowns. It also improves tooltips and adds insight legends. The release requires ClickHouse: Postgres-only installations are no longer supported, with 1.30.0 identified as the last supported version. The "All time" filter now considers events from 2015 onward, and Minute insight intervals are being removed due to parsing and performance concerns. Multivariate Feature Flags leave beta as a premium feature; Experimentation is in beta on PostHog Cloud. ### Source excerpt PostHog 1.32.0 makes it easier to find what you want in the Persons & Groups page, introduces vertical funnels and sets the stage for the launch of Experimentation! ## npm 7 is now Generally Available DevFeed: [npm 7 is now Generally Available]() Original publisher: [Read original article]() Author: phanatic Published: 2021-02-02T17:19:32Z Content type: release Language: en Sources: [GitHub Changelog]() Topics: [npm](), [npm packages]() Tags: [blog](), [blog-post](), [changes](), [cli](), [details](), [global](), [install](), [latest](), [npm](), [post](), [published](), [registry](), [run](), [version](), [want]() ### AI overview npm CLI 7 is Generally Available and published as `latest`, so `npm install --global npm` installs it by default. To install npm 6 instead, run `npm install --global npm@6`. The excerpt points to a separate blog post for details, including breaking changes. ### Source excerpt The npm CLI version 7 is now Generally Available. Please note, npm 7 is now published as latest to the npm registry and will be the default version installed when... ## Using Apollo Federation with Local Schemas DevFeed: [Using Apollo Federation with Local Schemas]() Original publisher: [Read original article]() Author: Patrick Arminio Published: 2020-10-06T16:05:38Z Content type: tutorial Language: en Sources: [Apollo Blog]() Topics: [Apollo Federation](), [GraphQL]() Tags: [api](), [apollo](), [apollo-federation](), [backend](