<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Tide Foundation on Medium]]></title>
        <description><![CDATA[Stories by Tide Foundation on Medium]]></description>
        <link>https://medium.com/@tidefoundation?source=rss-bba93caba371------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*SnKdoFIge4g7BT8JK3HLOw.jpeg</url>
            <title>Stories by Tide Foundation on Medium</title>
            <link>https://medium.com/@tidefoundation?source=rss-bba93caba371------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 21:59:21 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@tidefoundation/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What Emergent Authority Makes Possible: Beyond Cybersecurity]]></title>
            <link>https://levelup.gitconnected.com/what-emergent-authority-makes-possible-beyond-cybersecurity-269aefad3d98?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/269aefad3d98</guid>
            <category><![CDATA[economics]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[privacy]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Fri, 11 Sep 2026 15:26:14 GMT</pubDate>
            <atom:updated>2026-09-11T15:26:14.563Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*glqdAdix613OoObfR1Z2aQ.jpeg" /></figure><p>The New Stack recently spent some time with us exploring an idea we have been working on at Tide for years. Their headline got straight to it: <strong>“Why authority is the attack surface AI security keeps missing.”</strong> <a href="https://thenewstack.io/tide-raziel-emergent-authority/">Read The New Stack article</a></p><p>That sounds like a cybersecurity problem. It is. Cybersecurity is where the consequences become obvious first, but the implications of Emergent Authority reach much further.</p><p>Because if authority itself can be redesigned, something much bigger changes. We can start building systems, businesses and institutions that do not require somebody, somewhere, to hold dangerous amounts of power just to make useful things happen. That opens some interesting doors.</p><h3>First, what is authority?</h3><p>Start with a simpler question. What gives a digital system the power to act? That power is authority. It is what lets a server decrypt your information, an administrator enter a production system, an identity provider say you are you, software approve a transaction, an AI agent act on your behalf, or a company decide who can access what.</p><p>Today, that authority almost always lives somewhere. There is a key in a vault. A privileged account. A signing secret. A service account. A root administrator. Some person or machine eventually has enough power to make the important thing happen.</p><p>Cybersecurity has spent decades trying to protect whoever holds that power. We surround those holders with firewalls, vaults, MFA, monitoring and endless patching, but the basic bargain stays the same. Defenders have to get everything right, while an attacker needs only one useful mistake.</p><p>Tide was built to change that underlying architecture, not add another defensive layer around it. <a href="https://tide.org/whitepaper">With Emergent Authority</a>, consequential authority is not permanently held by a person, server, administrator, vendor or AI. Instead, Tide’s <a href="https://tide.org/howtideworks">decentralized Cybersecurity Fabric</a> distributes the cryptographic capability across independent nodes. When identity, policy, context and intent line up, those nodes collectively perform the permitted action. The complete underlying authority never needs to exist in one place.</p><p>We started with a simple way of describing it.</p><blockquote>Keys no one will ever hold.</blockquote><p>The more important idea hiding underneath is this:</p><blockquote>Power can be exercised without somebody possessing the power.</blockquote><p>That changes security. It may change quite a lot more.</p><h3>What Emergent Authority means for developers</h3><p>Developers have inherited an increasingly ridiculous job. Build quickly. Use hundreds of dependencies. Connect dozens of services. Manage identities, permissions, secrets and infrastructure. Keep everything patched. Configure everything correctly. Make sure no credential leaks. And now, do it while AI produces software faster than anyone can realistically inspect it. Meanwhile, an attacker only has to find one useful mistake.</p><p>That imbalance is why we built <a href="https://tide.org/tidecloak">TideCloak</a>, our Keycloak-compatible Identity, Immunity and Access Management layer. It brings Emergent Authority into familiar OIDC and OAuth application flows without asking developers to become cryptographers. More recently, we built <a href="https://tide.org/agent-plugins">Raziel</a>, Tide’s MCP server for AI coding agents, to identify concentrations of authority and guide agents through integrating TideCloak into a developer’s project.</p><p>But the bigger developer story is not another security tool. It is a different bargain:</p><ul><li>What if your application could be full of ordinary software bugs, yet those bugs could not reveal the signing key because the signing key was never there?</li><li>What if compromising your identity server did not automatically grant the attacker the power to forge identities?</li><li>What if an AI-generated application could interact with sensitive information without the backend becoming another place capable of reading all of it?</li></ul><p>This is what we mean by changing the architecture instead of endlessly defending the prize. The developer can get back to building. And that matters because when the cost of safely building something falls, we tend to build a lot more of it.</p><h3>Security is only the first-order effect</h3><p>Here is where things get more interesting. The most important consequence of safer cars was not that people worried less about crashing. Cars made mobility practical at enormous scale. That changed where people lived, how businesses operated, what could be transported, where people worked and what they did with their weekends. The safety improvement enabled everything downstream.</p><p>Emergent Authority opens the door to a similar second-order effect. The immediate outcome is that dangerous authority becomes harder to steal or misuse. The larger outcome is that we can start to cooperate without first having to trust somebody with dangerous power.</p><p>And trust is currently one of the great hidden constraints on what we can do together.</p><h3>AI could become a real delegate</h3><p>Consider AI agents. An assistant that tells you how to lower your electricity bill is useful. An agent that can read your private usage data, change providers, negotiate the contract, make the payment and update your accounts is much more useful. But now it needs the authority to perform those actions.</p><p>The more useful the agent becomes, the more frightening that becomes too. We do not particularly want a clever piece of software carrying reusable credentials to our identity, finances, company infrastructure and private information.</p><p>Emergent Authority creates another possibility. Give the agent the ability to get legitimate work done without giving it permanent possession of the authority behind that work. AI can move from advising people to genuinely working for them, while the infrastructure remains capable of saying what it may do, under which circumstances, and on whose behalf.</p><p>That could make sophisticated personal administration available to almost everyone. A person could have software continuously dealing with bills, travel, government paperwork, investments, insurance, subscriptions and business administration. Things wealthy people solve today with assistants, accountants, lawyers and family offices could increasingly become ordinary software capabilities.</p><p>Not because AI suddenly became trustworthy. Because trust became less necessary.</p><h3>Small companies could act much bigger</h3><p>The same thing happens to businesses. A five-person startup can already rent computing infrastructure that once required a multinational-sized data center. AI is now giving that same small team access to extraordinary cognitive capability. Authority remains one of the bottlenecks.</p><p>A major bank can hire a small specialist today, but giving that specialist meaningful access to sensitive systems creates risk. Companies keep work inside organizational boundaries partly because insiders are easier to govern than outsiders. Reduce that risk and the boundaries of the firm can become much more fluid.</p><p>A small company could safely assemble AI agents, contractors, specialist suppliers and infrastructure from around the world, giving each exactly the authority needed to contribute without turning each one into another dangerous insider. That means smaller companies can coordinate much larger amounts of capability.</p><p>Remote work becomes deeper than video meetings. Global specialists can participate in consequential work without being handed broad access. Tiny teams become viable competitors in industries where security and trust requirements currently favor giant incumbents. The productivity gain could dwarf the security gain.</p><h3>Valuable data could finally be used</h3><p>Some of humanity’s most valuable information is also information we are least willing to share: medical histories, genomic datasets, financial activity, industrial performance, infrastructure data and scientific research. There are good reasons for that.</p><p>A hospital might learn far more about a rare disease by combining information with hundreds of hospitals around the world. But doing so traditionally means somebody has to collect, control or gain access to an extraordinary pool of sensitive information. So much of the potential value remains stranded.</p><p>A world where useful operations can occur without one party receiving unilateral authority changes the equation. Hospitals could collaborate without creating a new holder of everyone’s medical records. Banks could collectively detect fraud without handing one organization complete visibility into every customer’s financial life. Companies could contribute information to shared AI or industry analysis without first surrendering it to another platform.</p><p>Privacy then stops being only about preventing something bad. It becomes an enabler of things we currently cannot safely do. More research. Better models. Better medicine. Better collective intelligence.</p><h3>We could get scale without always creating a ruler</h3><p>There is an even larger pattern here. Whenever humans want to coordinate something big, we tend to create something powerful to coordinate it:</p><ul><li>Banks coordinate capital.</li><li>Platforms coordinate digital markets.</li><li>Governments coordinate public services.</li><li>Large companies coordinate thousands of workers.</li></ul><p>Someone ultimately controls the infrastructure, the database, the money, the permissions or the keys. Scale tends to create custody, and custody tends to create concentrated power. That is not always bad. Often it is simply the only practical architecture we have had.</p><p>Emergent Authority suggests another possibility. What if a group could coordinate a shared resource without appointing one participant as its technical sovereign?</p><p>Communities could pool capital around local infrastructure without one person holding the master account. Companies that compete with one another could operate shared systems without one competitor controlling them. Governments that do not entirely trust one another could still cooperate around shared infrastructure because no country has unilateral control. New organizations could appear around a project and disappear when the project is finished, without first constructing an elaborate hierarchy of custodians.</p><p>The opportunity is not decentralization for its own sake. It is <strong>coordination at scale without custody at scale</strong>, an idea The New Stack picked up in its exploration of Tide. That is a much more useful goal.</p><h3>A lower trust economy could simply do more</h3><p>Trust is expensive. We rarely see the bill because it is buried across the economy. It appears in audits, insurance, compliance, supervision, escrow, bureaucracy, intermediaries and the countless things organizations choose not to do because the trust assumptions are too dangerous.</p><p>It also shapes our institutions. We favor large suppliers because we trust them more than unknown startups. We keep data siloed because sharing it creates liability. We keep humans supervising processes that software could perform because giving software authority feels reckless.</p><p>We also wrap critical access in layers of infrastructure because the intermediary itself has to be trusted. Applying Emergent Authority to manage infrastructure access in the form of <a href="https://tide.org/blog/vibe-coded-pam">KeyleSSH</a> shows what changes when that intermediary becomes an oblivious proxy. It can broker remote access without holding the keys or seeing the session, allowing much of the surrounding security infrastructure to disappear with it. The result is not only stronger security, but simpler infrastructure and better performance.</p><p>Lower the amount of trust required, and some of those constraints move:</p><ul><li>More work can cross company boundaries.</li><li>More capital can form around smaller opportunities.</li><li>More assets can be shared.</li><li>More information can become useful.</li><li>More organizations and individuals can adopt powerful AI with confidence.</li><li>More strangers can collaborate.</li><li>More small organizations can compete with large ones.</li></ul><p>The important outcome is not a world where nobody trusts anybody. Quite the opposite. It’s a world where cooperation no longer requires making yourself dangerously vulnerable to whoever sits on the other side, or taking on liabilities you never wanted to carry in the first place.</p><h3>Cybersecurity is where this starts</h3><p>We built Emergent Authority because cybersecurity has an authority problem. Traditional systems place enormous power inside extractable things, then spend enormous effort protecting those things. A single credential, administrator, signing key or service can become the path through which an otherwise sophisticated system collapses.</p><p>Tide’s work has been about removing that structural weakness. The authority behind authentication, authorization, decryption and privileged access can be kept beyond the unilateral reach of the systems and people using it. That alone is worth pursuing.</p><p>The more we work with the idea, the clearer it becomes that cybersecurity is only where Emergent Authority begins.</p><p>Human progress has always depended on better ways to coordinate with people beyond those we personally know and trust. Money helped strangers trade. Contracts helped strangers make commitments. Companies helped strangers pool capital. The internet helped strangers exchange information and build together across the planet.</p><p>Perhaps the next step is making it possible for strangers, companies, machines and institutions to cooperate deeply without requiring any of them to hold disproportionate power over the others.</p><p>That is the opportunity behind Emergent Authority. A safer internet would be useful. A world capable of coordinating vastly more human and machine capability without concentrating vastly more power is the real opportunity.</p><h3>About the author</h3><p><a href="https://tide.org/">Tide Foundation</a> is the team behind Emergent Authority, an approach to designing digital systems so that consequential power does not need to be held by any single person, system or organization.</p><p>Tide researches and builds the cryptographic infrastructure behind the idea, with work spanning identity, data protection, privileged access and AI. If you’d like to go deeper into how Emergent Authority works technically, the research and architecture are available in Tide’s <a href="https://tide.org/whitepaper">Emergent Authority whitepaper</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=269aefad3d98" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/what-emergent-authority-makes-possible-beyond-cybersecurity-269aefad3d98">What Emergent Authority Makes Possible: Beyond Cybersecurity</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[HIPAA Security for HealthTech: How to Protect ePHI from a Single Compromise]]></title>
            <link>https://itnext.io/hipaa-security-for-healthtech-how-to-protect-ephi-from-a-single-compromise-97b94d827d8a?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/97b94d827d8a</guid>
            <category><![CDATA[gdpr]]></category>
            <category><![CDATA[hipaa-compliance]]></category>
            <category><![CDATA[identity-management]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[privacy]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Thu, 10 Sep 2026 08:02:31 GMT</pubDate>
            <atom:updated>2026-09-10T08:02:31.073Z</atom:updated>
            <content:encoded><![CDATA[<h3>IPAA Security for HealthTech: How to Protect ePHI from a Single Compromise</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4gRT1dVMu2ShE4epIoWiGQ.jpeg" /></figure><blockquote><strong>TL;DR: </strong>If your platform creates, receives, maintains or transmits health information on behalf of a Health-related entity, HIPAA may apply to you. Start with the data you handle and the obligations that apply, then test whether a single compromised component could expose ePHI.</blockquote><p>The US Department of Health and Human Services (HHS) is moving toward more explicit, testable cybersecurity requirements. Health platforms will need to strengthen MFA, encryption, asset maps, testing and audits. That raises the bar, but it does not remove the core catastrophic failure path.</p><p>If one trusted component can impersonate a user, change policy or recover plaintext, one compromise can still become an ePHI breach. That is the gap worth closing before the next rule, or attacker, forces another policy redesign.</p><p>This guide covers what counts as PHI and ePHI, when HIPAA applies to a software platform, and how to assess the controls protecting that data.</p><h3>What is Protected Health Information (PHI)?</h3><p>Protected Health Information (PHI) is broader than most engineering teams assume.</p><p>Protected Health Information is any information about a person’s health status, health care, or payment for health care that’s created, received, stored, or transmitted by a HIPAA-covered entity or its business associates, and that can be linked back to a specific individual.</p><p>It’s a subset of personally identifiable information, but specifically the kind tied to health records or health services.</p><p>HIPAA defines 18 identifiers that turn ordinary information into PHI the moment it’s connected to health data:</p><ol><li>Names,</li><li>Geographic details smaller than a state,</li><li>Dates other than year,</li><li>Phone</li><li>Fax numbers,</li><li>Email addresses,</li><li>Social Security numbers,</li><li>Medical record numbers,</li><li>Health plan numbers,</li><li>Account numbers,</li><li>License numbers,</li><li>Vehicle identifiers, serial numbers, such as licence plate numbers</li><li>Device identifiers, serial numbers</li><li>Uniform Resource Locators (URL),</li><li>Internet Protocol (IP) addresses,</li><li>Biometric identifiers,</li><li>Full-face photos,</li><li>Any other unique identifying code.</li></ol><h3>What is not Protected Health Information (PHI)</h3><p>Data that can’t identify a patient, employment records held by a healthcare organization acting as an employer rather than a provider, and information about someone who has been deceased for more than 50 years is not considered PHI.</p><p>Encryption alone does not de-identify a dataset, if the decrypted data still contains the original identifiers.</p><p>De-identified or anonymized data (with the above 18 identifiers stripped and a low re-identification risk validated by a statistician) conforms to HIPAA’s Safe Harbor de-identification requirements and falls outside PHI, which is why it can be used for research in ways identified data can’t.</p><h3>What is Electronic Protected Health Information (ePHI)</h3><p>Electronic PHI, or ePHI, is the same thing as PHI, when it’s created, stored, or transmitted electronically, which covers most modern health technology, such as:</p><ul><li>Patient portals,</li><li>Scheduling apps,</li><li>Electronic Health Record (EHR) systems,</li><li>Even a support ticket that references a patient by name.</li></ul><p>If your software product provides a service to a covered entity and touches any of that data, you’re very likely a business associate under HIPAA regardless of whether you think of yourselves as a health tech company, and you carry similar legal exposure.</p><p>PHI is what triggers HIPAA’s Privacy Rule (governing use and disclosure) and Security Rule (governing technical, physical, and administrative safeguards), and it’s what turns a routine data exposure into a reportable breach with real penalties attached.</p><h3>What does this mean for HealthTech companies?</h3><p>The Health Insurance Portability and Accountability Act of 1996 (HIPAA), mandates that Covered Entities implement security controls to protect the confidentiality, integrity, or availability of that information.</p><p>Covered entities include hospitals, doctors, health plans, insurers, clearinghouses, and their business associates — such as HealthTech companies or software companies who build software for the Health industry.</p><h3>Does HIPAA apply to your platform?</h3><p>HIPAA does not apply to all health data. A software vendor may be a business associate when it creates, receives, maintains or transmits PHI on behalf of a covered entity or another business associate.</p><p>For example, a database entry that just says “patient took drug X” isn’t PHI on its own, but the moment it’s linked to a name, an email, or even just an IP address, it becomes PHI.</p><p>Patient portals, cloud storage, remote monitoring and provider-supplied apps can fall within that definition. Some direct-to-consumer wellness apps may not, although other laws can still apply.</p><p>A practical self-assessment:</p><ol><li>Are we serving a covered entity or business associate?</li><li>Do we handle individually identifiable health information for that customer?</li><li>Are we doing so on the customer’s behalf?</li></ol><p>If the answer is yes, assess your business associate agreements, subcontractors and the full Privacy, Security and Breach Notification obligations against the regulations — not only your technical safeguards.</p><h3>The cost of exposing patient data</h3><p>Healthcare and public health recorded <strong>460 ransomware complaints</strong> in the <a href="https://www.ic3.gov/AnnualReport/Reports/2025_IC3Report.pdf#page=16">FBI’s 2025 IC3 report</a>, more than any other US critical infrastructure sector.</p><p>Healthcare also has the highest average breach cost of any industry in <a href="https://www.ibm.com/think/topics/data-breach">IBM’s 2026 figures</a>.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/adab252e2efe01691bbe9ebf4df77b43/href">https://medium.com/media/adab252e2efe01691bbe9ebf4df77b43/href</a></iframe><p><em>Source: </em><a href="https://www.ibm.com/think/topics/data-breach"><em>IBM, 2026</em></a><em>. Healthcare’s average is approximately 33% higher, calculated from the published figures.</em></p><p><strong>For HIPAA violations,</strong> civil penalties are tiered by culpability, not by company size.</p><p>As of the January 2026 update:</p><ul><li>Tier 1 (violation you were unaware of and couldn’t reasonably have avoided) runs $145 to $73,011 per violation.</li><li>Tier 2 (reasonable cause, not wilful neglect) is $1,461 to $73,011 per violation.</li><li>Tier 3 (wilful neglect, but corrected within 30 days) is $14,602 to $73,011 per violation.</li><li>Tier 4 (wilful neglect, not corrected) is $73,011 up to $2,190,294 per violation.</li></ul><p>All four tiers share the same annual cap of $2,190,294 for multiple violations of an identical requirement.</p><p>Separately, the Department of Justice can pursue criminal charges for knowing violations, ranging from up to $50,000 and a year in prison for a basic knowing violation, up to $250,000 and 10 years for violations committed for commercial advantage, personal gain, or malicious harm.</p><h3>What the Data Protection Rule requires, and what’s about to get stricter</h3><p>The <a href="https://www.hhs.gov/hipaa/for-professionals/security/laws-regulations/index.html">current Security Rule</a> requires reasonable and appropriate administrative, physical and technical safeguards. ‘Addressable’ coverage is not optional: it requires implementation, an appropriate alternative or documented reasoning. Its technical safeguards cover the following:</p><ul><li>Access control (unique user identification, emergency access, automatic logoff),</li><li>Audit controls (recording and examining system activity),</li><li>Integrity controls (protecting data from improper alteration),</li><li>Person or entity authentication (verifying who’s actually asking), and</li><li>Transmission security (protecting data in transit).</li></ul><p>The <a href="https://www.hhs.gov/hipaa/for-professionals/security/hipaa-security-rule-nprm/factsheet/index.html">2025 proposal</a> would go further once in effect, including asset inventories and ePHI maps, MFA, encryption at rest and in transit with limited exceptions, vulnerability scanning, penetration testing, annual audits and stronger business-associate verification.</p><p>The direction is clear: controls should be specific, testable and evidenced. Yet a control can be present and still fail if a “trusted” component can bypass it.</p><h3>The “one-compromise” test</h3><blockquote><strong>Ask: </strong>If any single administrator, identity service, application server, key system or vendor is compromised, can it expose ePHI?</blockquote><p>MFA does not prevent a compromised identity provider from issuing a token, or malware from stealing a live session. Encryption at rest and in transit does not stop plaintext appearing on an application server during use. Logs do not independently prove that a change was authorised if the same system can alter policy and evidence. A break-glass made by a single account is nothing short than a bypass.</p><p>For each scenario, record what the compromised component can do and which independent control, if any, would still protect the data. Use the same questions when assessing your own platform or a vendor’s product.</p><h3>How TideCloak changes the failure modes which result in PHI Data Breaches</h3><p>TideCloak is an Identity, Access, and Authority Management (IAAM) solution built from the ground up with different thinking. It is designed to move critical authority out of the reach of any single administrator, server, cloud provider or vendor (itself included).</p><p>The distinction is best understood by the failure modes it is designed to remove:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/f6c2d4d6769bdfba2c28bbbdf9328538/href">https://medium.com/media/f6c2d4d6769bdfba2c28bbbdf9328538/href</a></iframe><p><em>Architecture mapping only. HIPAA obligations depend on the entity, use case, deployment and operational safeguards.</em></p><h3>Where to start</h3><p>Start with the applicability self-assessment, then run the one-compromise test against your current design. Record where ePHI appears in plaintext, who can grant access, and whether emergency access or policy changes depend on a single account.</p><p>HIPAA’s next requirements will help health platforms close known gaps. The more durable architectural question is whether ePHI stays protected when a trusted component does not.</p><h3>About the author</h3><p><a href="https://tide.org/">Tide Foundation</a> develops <a href="https://tide.org/tidecloak">TideCloak</a>, the identity and authority architecture discussed in this article. Its research focuses on designing systems where the compromise of any single administrator, service or infrastructure provider is insufficient to expose sensitive data.</p><p>For developers interested in evaluating the approach themselves, TideCloak is available to run in a development environment through the technical <a href="https://docs.tide.org/get-started/Tidecloak%20Quickstart">quickstart</a>.</p><p><em>This article provides general information, not legal advice. Consult qualified counsel about how HIPAA and other health-data laws apply to your organization and product.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=97b94d827d8a" width="1" height="1" alt=""><hr><p><a href="https://itnext.io/hipaa-security-for-healthtech-how-to-protect-ephi-from-a-single-compromise-97b94d827d8a">HIPAA Security for HealthTech: How to Protect ePHI from a Single Compromise</a> was originally published in <a href="https://itnext.io/">ITNEXT</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What your next cyber insurance renewal will demand]]></title>
            <link>https://itnext.io/what-your-next-cyber-insurance-renewal-will-demand-23fc3c61e2ab?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/23fc3c61e2ab</guid>
            <category><![CDATA[insurance]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cryptography]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Mon, 29 Jun 2026 08:12:33 GMT</pubDate>
            <atom:updated>2026-06-29T08:12:33.355Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/998/1*2Y2ZqCzBWeCZEoEs2Qlmxg.jpeg" /></figure><p><strong>Cyber insurers are moving past annual declarations and point-in-time evidence. The market is shifting toward controls that can be cryptographically verified, continuously enforced, and tied to real loss reduction.</strong></p><p>Cyber insurance is changing the way security programs are judged.</p><p>For years, renewal was built around declarations. Organizations said they had MFA. They said backups were immutable. They said privileged access was controlled. They said third-party access was reviewed.</p><p>Insurers priced the risk on those answers, the broker’s context, and whatever evidence could be attached.</p><p>But a questionnaire answer is not a control. A screenshot is not a control. A policy document is not a control.</p><p>They are claims about a control.</p><p>Cyber insurance is now moving from claims to proof. Not more paperwork. Not better screenshots. Proof.</p><p>A provable control is one that can be cryptographically verified. It replaces blind trust with math. Instead of asking an insurer, broker, board, or customer to believe that a control exists, it lets the system prove that the control is enforced and that a critical action cannot occur unless the required conditions are met.</p><p>That shift is good for security teams and good for insurers. When a control is provable, it removes uncertainty from the security program. The same uncertainty is removed from the underwriting conversation. The risk is clearer for both sides.</p><p>The companies that become easiest to insure will not be the ones with the longest control inventories. They will be the ones whose most important controls are no longer a matter of opinion.</p><h4>The old model was built on weak signals</h4><p>Cyber insurance tried to apply traditional insurance models to a market without centuries of actuarial history.</p><p>That meant questionnaires, declarations, evidence packs, and underwriting judgment had to carry too much weight. In practice, this created a market built on weak signals.</p><p>An organization could answer “yes” to MFA while leaving bypass paths open. It could claim immutable backups while allowing the same compromised admin path to disable recovery. It could point to privileged access controls while standing authority remained concentrated in admin accounts, service accounts, vendors, API keys, or security tools.</p><p>The statement may be technically true. The risk may still be unacceptable.</p><p>That is the gap now being closed.</p><p>Underwriters are no longer only asking whether a control exists. They are asking whether the control changes the outcome of an attack.</p><h4>Provable controls change the underwriting conversation</h4><p>An attested control says, “We have this.”</p><p>A provable control says, “This attack path is closed.”</p><p>That is a very different risk statement.</p><p>Attestation depends on trust, interpretation, and point-in-time evidence. Provable control depends on enforcement that can be verified. In Tide’s view, the strongest version of this is cryptographic verification. Critical authority is not simply handed to one person, account, vendor, or system. It only becomes usable when the right conditions are met.</p><p>This changes cybersecurity architecture, but it also changes insurance.</p><p>If a security team can prove that a compromised administrator cannot destroy recovery, the underwriter does not have to price that risk as if recovery depends on hope. If a company can prove that no single vendor can exercise unchecked control inside its environment, the insurer does not have to treat that dependency the same way as blind third-party trust.</p><p>Security teams have been trying to remove single points of failure for years. Underwriters are now being forced toward the same conclusion because the same weaknesses that create breaches also create large claims.</p><p>Take the worry away from the security team and you take it away from the underwriter. Take it away from the underwriter and the commercial value of better security becomes easier to recognize.</p><h4>MFA is no longer a yes-or-no question</h4><p>MFA is the simplest example.</p><p>The market started with “Do you have MFA?” Then it moved to “Do you have MFA everywhere that matters?”</p><p>Now the real question is whether MFA can be bypassed.</p><p>Many of the breaches that made headlines in recent years did not break MFA. They went around it. Attackers used stolen session tokens, hijacked cookies, push fatigue, help desk manipulation, compromised endpoints, and weaknesses in identity workflows.</p><p>The same problem shows up in security products themselves. Authentication and access-control bypasses continue to be found in critical software, including identity, PAM, and other cybersecurity platforms. That should make the point obvious. A control that depends on a single trusted path can become the thing attackers route around.</p><p>For an underwriter, these are different risks.</p><p>One company may have MFA deployed but still allow a stolen session or bypassed workflow to reach sensitive systems. Another may have an architecture where access cannot proceed unless the correct authentication has happened in the correct context.</p><p>Those companies should not be priced the same.</p><p>The useful test is simple enough.</p><p>Could a sensitive or privileged action still happen without fresh, valid, context-bound authentication?</p><p>If the answer is yes, the organization has an exposure. If the answer is no and that fact can be verified, it has a provable control.</p><h4>Standing authority is the real enemy</h4><p>The single biggest question behind many cyber losses is not whether the company had tools. It is who, or what, held authority when the breach happened.</p><p>Standing authority is the always-on power sitting inside admin accounts, service accounts, vendor connections, API keys, cloud roles, and AI agents. It is the power to read data, change systems, disable controls, move laterally, deploy code, or destroy recovery.</p><p>Most security programs try to manage that power by watching it more closely. They vault passwords. They rotate credentials. They log sessions. They require approvals.</p><p>Those steps help, but they often preserve the dangerous premise. Somewhere in the environment, power still exists in a concentrated form, waiting to be stolen or misused.</p><p><a href="https://tide.org/blog/cyber-immunity-emergent-authority-ai-era">Tide’s view is different</a>. Authority should not sit statically in one place. It should be emergent.</p><p>Emergent authority means power only comes into existence when the right parties, context, and conditions align. Outside that moment, there is no complete key sitting inside an account, vault, tool, or vendor for an attacker to steal.</p><p>That changes the loss profile at the root.</p><p>A compromised administrator who holds unilateral power can be catastrophic. A compromised administrator who never actually holds the full authority to act alone is a fundamentally different risk.</p><p>That distinction is exactly what cyber insurance needs to recognize.</p><h4>Backups must survive admin compromise</h4><p>Backups show why tick-box security fails.</p><p>“Do you have backups?” stopped being a useful question years ago. Ransomware operators now plan around the recovery path. They look for backup consoles, admin credentials, retention settings, replication paths, and recovery dependencies.</p><p>Immutability helps, but immutability is not enough.</p><p>If the same compromised admin path that controls production can disable, reconfigure, delete, or age out the backup environment, then immutability is just a setting with an off switch.</p><p>The real test is blunt.</p><p>If ransomware lands tonight and the admin tier is fully compromised, does recovery still exist?</p><p>A strong architecture separates recovery authority from production authority. Owning one should not quietly grant ownership of the other. The recovery path should be provably separate, not merely documented as separate.</p><p>That gives the insurer a better signal than “we have backups.”</p><p>It says a full compromise of production administration does not grant the ability to delete, disable, or corrupt recovery.</p><p>That is the difference between evidence and proof.</p><h4>Supply chain risk is authority risk</h4><p>Supply chain risk becomes much clearer when viewed through authority.</p><p>Your identity provider, remote access vendor, backup platform, endpoint tool, cloud provider, and software supply chain are not outside your threat model. Many of them sit inside your control plane.</p><p>If a vendor can push code, administer identities, access data, change security settings, or disable protections, that vendor holds real authority inside your environment.</p><p>For one company, the task is to understand what each supplier can do and whether that power can be constrained. For insurers, the concern is larger. If many insured organizations depend on the same concentrated vendors, one supplier compromise can create correlated losses across a portfolio.</p><p>That is the aggregation risk insurers fear.</p><p>Architectures that reduce concentrated vendor authority are therefore not only better security. They are better insurance risks.</p><p>If no single vendor, tool, or administrator can exercise unchecked authority, the company has reduced the chance that one upstream compromise becomes a catastrophic downstream claim.</p><h4>AI makes this urgent</h4><p>AI increases the cost of weak authority models.</p><p>AI writes code, generates scripts, suggests configurations, processes sensitive data, and increasingly acts through connected tools. AI agents will not only recommend actions. They will execute them.</p><p>Insurers do not yet have a complete model for pricing AI risk. Nobody does.</p><p>But the control principle is already clear. Reduce the authority any one AI system can exercise, and make every agent accountable to a defined owner, role, and policy.</p><p>An AI agent should not become a magic super-admin. It should operate inside bounded authority, just like a person should. Preferably, that authority should not be static power sitting in the agent at all. It should emerge only when the right conditions are met.</p><p>That is how organizations can adopt AI without turning every automation into a new insurance exposure.</p><h4>How to prepare for the next renewal</h4><p>The next renewal should be used to separate claims from proof.</p><p>For each major control, security leaders should be able to show three things.</p><ul><li>Where the control is deployed.</li><li>What loss scenario it is meant to prevent.</li><li>Whether the control can be verified as enforced.</li></ul><p>That framing is more useful than another long evidence pack.</p><p>Instead of saying the organization has privileged access management, show that no single administrator, account, or vendor can unilaterally exercise critical authority.</p><p>Instead of saying the organization has immutable backups, show that production compromise does not compromise recovery.</p><p>Instead of saying AI use is governed, show that AI agents have bounded authority and accountable owners.</p><p>The language is shifting from control existence to loss prevention.</p><p>That is the language underwriters understand.</p><h4>Cyber insurance is becoming a proof market</h4><p>Cyber insurance is moving away from annual declarations and toward continuous assurance.</p><p>Once control integrity can be proven live, the annual questionnaire becomes less important. Security posture becomes a signal. That signal can be monitored, priced, and recognized.</p><p>The old model created too much ambiguity. The insured declared a control. The insurer priced the risk. Months later, after an incident, everyone discovered whether the control was real enough to matter.</p><p>That model is being replaced because it has to be.</p><p>Insurers need clearer signals. Brokers need stronger ways to distinguish clients. Security teams need their real risk reduction to be recognized commercially. Boards need to know whether security spend is reducing loss, not just satisfying a form.</p><p>The companies that adapt will be easier to understand, easier to price, and easier to insure.</p><p>The winners will be the organizations whose controls can be verified by design.</p><blockquote>The ideas in this piece are the principle behind Tide’s collaboration with <a href="https://www.insurancenews.com.au/insurtech/guaranteed-cybersecurity-wins-lloyd-s-backing">Lloyd’s-backed underwriter Becco</a>, which links provable controls to preferential cyber insurance coverage. It creates a commercial path for recognizing a better class of cyber risk, where security controls can be verified rather than merely asserted.</blockquote><blockquote>We recently contributed to <a href="https://www.helpnetsecurity.com/2026/06/25/cyber-insurance-controls-video/">Help Net Security’s video series</a> on what cyber insurance renewals are now demanding. This piece sets out the shift security leaders need to prepare for.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=23fc3c61e2ab" width="1" height="1" alt=""><hr><p><a href="https://itnext.io/what-your-next-cyber-insurance-renewal-will-demand-23fc3c61e2ab">What your next cyber insurance renewal will demand</a> was originally published in <a href="https://itnext.io/">ITNEXT</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I vibe-coded a PAM]]></title>
            <link>https://levelup.gitconnected.com/i-vibe-coded-a-pam-781589d2ade2?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/781589d2ade2</guid>
            <category><![CDATA[authentication]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Fri, 05 Jun 2026 15:45:59 GMT</pubDate>
            <atom:updated>2026-06-05T15:45:59.365Z</atom:updated>
            <content:encoded><![CDATA[<h3>The only way to securely “vibe code” is to take away the keys</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0BehyQdw7IQ3m73xvDzHWQ.jpeg" /></figure><p><em>This is the written-up version of a talk delivered by Sasha Le, Senior Engineer, </em><a href="https://tide.org/"><em>Tide Foundation</em></a><em>, at BSides Gold Coast. If you’ve read our earlier posts on </em><a href="https://tide.org/blog/cyber-immunity-emergent-authority-ai-era"><em>Emergent Authority</em></a><em> and </em><a href="https://tide.org/blog/keylessh"><em>KeyleSSH</em></a><em>, this is a more technical walkthrough.</em></p><p>My name is Sasha, and I vibe-coded a PAM.</p><p>That should worry you. A Privileged Access Management system is, by design, the single most concentrated pile of risk in an organization. It centralizes every privileged credential you own into one place. It is the keys to the kingdom. So when an attacker got hold of an API key from PAM vendor BeyondTrust last year, the tool that exists to <em>protect</em> privileged access became the single point of catastrophic failure for at least 17 organizations, including the US Treasury.</p><p>But I can’t blame them. They’re fighting the same losing battle as all platform developers.</p><p>That’s the paradox we keep coming back to. A PAM is built to guard the kingdom, but to do its job we hand it every key in the kingdom. Building one badly, by hand, over a few weekends, should be a recipe for disaster.</p><p>It doesn’t have to be, and the reason why is the whole point of this post.</p><h4>The problem doesn’t get solved. It moves.</h4><p>Start with the simplest possible remote access, a password SSH login. Your console opens a tunnel to the server, you type a username and password, the server checks them against its own copy, and you’re in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*PeokOd18fCCkt_7Z.gif" /></figure><p>Look at what that costs you. The credentials have to map one-to-one to an account on the box. An admin has to remember to kill that account when the person leaves. And a copy of the password lives on the server, so anyone with read access can impersonate the user.</p><p>So best practice says don’t use passwords, use PKI. Now the secret never leaves the user, authentication is zero-knowledge, and one account can be authorized by multiple keys. Better, genuinely better. But the problems didn’t disappear. They <em>moved</em>. Now the user has private keys to manage. The admin still has to pull those keys when someone leaves, rotate them on a schedule, and keep a clean mapping of every public key to every user.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*Q94709LgaGdNwcbr.gif" /></figure><p>That adds plenty of admin overhead, but it doesn’t fundamentally change the security situation. A key stored anywhere is a key exposed somewhere. Every honest attempt to secure SSH does the same thing. It relocates the secret to a slightly safer-looking box, a bastion, a vault, an HSM, and calls it progress.</p><p>The only way to win is to stop relocating the secret and remove the thing the attacker is reaching for.</p><h4>Emergent Authority, in one move</h4><p>The principle underpinning KeyleSSH is what we call Emergent Authority. The authority to perform a sensitive action sits completely outside the reach of any system or person using it, and emerges <em>just in time</em>, only because the right context and conditions have been met.</p><p>So the system’s sensitive data, identities and access rights get locked with keys that no one, not us, not you, not your admins, will ever hold as a whole. Those keys are generated and operated in fragments across the <a href="https://tide.org/howtideworks">Tide Cybersecurity Fabric</a>, a decentralized network of nodes that perform partial operations on secrets without ever reconstructing them. Not even inside a Trusted Execution Environment. The signature is assembled from operating on those distributed pieces, and the key itself is never expressed anywhere.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*6GfKVC1-dihNkcaX.gif" /></figure><p>Here is how that plays out in practice. When a sensitive operation is needed, KeyleSSH sends a simple API request out to the Fabric nodes. Each node, uses its own key fragment to perform just a small part of the operation, whether that’s authentication, decryption, or signing, and hands back nothing more than its slice of the result.</p><p>What comes back from each node is a meaningless puzzle piece. Individually they reveal nothing. Only an authorized user can assemble the pieces into a meaningful result, which leaves everyone else in the blind and the key hidden forever, even from us. In the case of an SSH session, those partial signatures combine into a single valid Ed25519 signature, and the private key behind it is never reassembled anywhere.</p><p>None of this is magic, it’s verifiable mathematics. The keys are born in fragments and stay that way, they never come together into a whole, anywhere. If you’re familiar with threshold cryptography, secure multi-party computation, or zero-knowledge proofs, you’ll recognize the shape of it. We call the framework Ineffable Cryptography, because the secret is never expressed, not even for an instant.</p><p>We’ve covered that cryptography in more detail elsewhere, so I won’t relitigate it here. What I want to walk through is what it actually <em>does</em> to a live SSH session, because that’s where it stops being a nice diagram and starts being load-bearing.</p><h4>A session, start to finish</h4><p><strong>Login.</strong> The user logs into the KeyleSSH web app exactly like any browser login they’ve ever done. Username, password, enter. The interesting part is what didn’t happen. No password left the device. There’s no password database, and no server, not even during verification, ever sees the password. Authentication is a zero-knowledge proof against the Fabric, and the permissions granted for the session are fingerprinted to the device so they can’t be replayed elsewhere.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*NvAIjHIIoRxkyN6p.gif" /></figure><p><strong>What you gain</strong></p><ul><li>No user credentials to protect</li><li>No key management, for the user or the security team</li><li>No replayable session token to steal</li><li>No change to the user experience</li><li>One management system, good for compliance</li><li>Zero-knowledge, decentralized authentication</li><li>Authentication that can’t be bypassed</li></ul><p><strong>Account assignment.</strong> Once authenticated, KeyleSSH maps the user to a pre-assigned SSH account representing their roles on the target server, and certifies that assignment for one ephemeral session using an organization master key that, again, sits outside anyone’s reach. To stop a compromised KeyleSSH server or a rogue admin from quietly handing access to whoever they like, that assignment is only issued once a quorum of admins has cryptographically signed off on it. No single server and no single admin can forge it. This isn’t application logic that a root admin can patch out. It’s a constraint enforced by the signing network itself. If the threshold isn’t met, the signature mathematically never forms.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*WMWHgSlcq_YU9h1W.gif" /></figure><p><strong>What you gain</strong></p><ul><li>Centralized management with no central point of vulnerability</li><li>Decentralized security</li><li>No god-like admin</li><li>An authorityless platform: rooting it doesn’t get you the authority</li></ul><p><strong>The tunnel.</strong> Now the session builds an end-to-end secure channel. HTTPS from the browser to the KeyleSSH proxy, then SSH from the proxy onward to the server. The whole thing is encrypted with an ephemeral session key that only the user’s terminal and the server ever hold. To trust the connection, the server checks exactly one thing, the single public key of KeyleSSH’s master key, and then honors the signed account assignment for that session only.</p><p>Here is the part worth sitting with. The proxy is an oblivious converter. HTTPS in, SSH out. It cannot see into the tunnel and it cannot touch it. Breach this PAM and you find an empty shell. There is no vault to aim at, because the authority to open the session was never in the proxy to begin with. Bribe your way into it and the best you get is the compromised user’s already-scoped access, with automatic throttling the moment behavior looks off. And it looks off quickly, because every action requires the Fabric’s approval. Nothing bypasses policy, and the Fabric keeps an immutable audit trail, so accountability is baked into every action. Even a disgruntled insider has cause to pause.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*XYvHmvpev51uamAf.gif" /></figure><p><strong>What you gain</strong></p><ul><li>No blindly trusted anchor</li><li>End-to-end verifiability, for both the user and the server</li><li>Zero-knowledge proof of authorization, with no secrets</li></ul><p>That’s the difference between a PAM and an <em>oblivious</em> proxy. One holds the kingdom’s keys and tries very hard not to lose them. The other never holds them, so there is nothing to lose.</p><h4>Making the internet as oblivious as the proxy</h4><p>Everything so far is already a hermetic model for SSH. But at scale there’s a question we haven’t reached yet. What happens when there’s an internet between the two ends?</p><p>Exposing an SSH server to the internet is its own category of pain. You’re opening or forwarding a port, standing up firewalls, issuing an SSL certificate, managing ACLs tied to the user’s source address, and then maintaining all of it, forever. Beyond the operational headache, every one of those seams is a place a new vulnerability can open up. Most “secure access” products spend the bulk of their complexity here, and most breaches find their foothold here.</p><p>KeyleSSH skips the whole thing. It can be hosted anywhere the user can reach it, including as a cloud-hosted service, and you drop a small <strong>Punch’d gateway</strong> inside the server’s network. It’s the same oblivious proxy as before, with one addition. It punches an <em>outbound</em> hole through the router out to the KeyleSSH infrastructure, the same idea as a Zscaler or a Cloudflare tunnel, except you don’t have to trust it. No inbound ports. No firewall choreography. No public certificate. A session request from the user now has somewhere to land, the gateway’s waiting outbound connection.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*86VQenYnVr7_Cwy4.gif" /></figure><p><strong>What you gain</strong></p><ul><li>Global cloud reach and scale</li><li>A Zscaler you don’t have to trust</li><li>Removed infrastructure: no firewalls, no port scanning</li><li>Simplified endpoint provisioning and maintenance</li><li>Topological flexibility, whether that’s out-of-band, DMZ, behind a VPN, or even air-gapped</li></ul><p>From there the two ends negotiate a direct, point-to-point tunnel. It runs over WebRTC’s UDP transport, and the gateway converts it back to plain SSH at the server edge. The tunnel <em>inside</em> stays end-to-end encrypted with keys only the user’s terminal and the server ever hold. Which means the traffic doesn’t even route through KeyleSSH. And in the case where it does, say the Punch’d gateway is fully compromised, it is still completely blind to everything flowing through it. An oblivious man-in-the-middle by construction, not by policy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*-fgpGOEe2bKlV8yJ.gif" /></figure><p><strong>What you gain</strong></p><ul><li>Point-to-point performance and security</li><li>Auto-discovery that just finds the path</li><li>A man-in-the-middle that’s entirely oblivious</li></ul><h4>It isn’t only SSH</h4><p>I started with SSH because a PAM is one of the hardest, highest-stakes things you could pick. But the model isn’t specific to SSH. It’s specific to <em>authority</em>, and authority is everywhere.</p><p>So the same edges now carry more protocols. There’s a secure reverse web proxy for internal sites. There’s browser-based remote desktop, where the RDP server authenticates with a JWT instead of a password, far less to attack. And there’s network-layer VPN access into internal infrastructure. Each one works the same way. The proxy converts protocols at the edge and stays blind to the contents, while the authority lives in the Fabric, not the code. We’ll dig into each of these in their own posts soon.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*xJam7gXyHq_FRvp3.gif" /></figure><p><strong>What you gain</strong></p><ul><li>Keyless SSH</li><li>Point to Point RDP</li><li>Trustless VPN</li><li>Secure reverse web proxy for internal sites</li></ul><h4>Trial by fire: the ceiling on secure vibe coding</h4><p>Now, back to where I started. If you wanted to know whether vibe coding can ever be secure, a PAM is the hardest place to ask the question. The entire platform, SSH, RDP, VPN, the reverse proxy, all of it, I vibe-coded in three weekends. So I can promise you it is full of horrific bugs.</p><p>None of them can touch the cryptography at the edges.</p><p>The authority simply isn’t in the code to steal. There’s no key in memory, no secret in a config, no vault behind the app. So the software has exactly two states. It works securely, or it breaks completely. It <strong>fails closed.</strong> A bug <em>can’t</em> leak the kingdom, because the kingdom’s keys were never in the building. Neither can malicious code, for that matter, and that distinction is getting more important by the month as supply-chain attacks through dev repos become routine. A poisoned dependency that would hand an attacker the keys in a traditional PAM finds the same empty shell everyone else does.</p><p>This is what changes the math for the next few years. Frontier AI models <em>will</em> find the vulnerabilities in your software, faster than any human review can keep up, across the 85% of production code that’s now AI-generated. Your admins <em>will</em> eventually be socially engineered. Those aren’t risks you patch your way out of. They’re virtual certainties.</p><p>The only durable answer is to make the breach inconsequential. To build systems where the worst day still yields nothing, because the authority an attacker is reaching for was never there for the taking. That’s cyber immunity.</p><h4>See how it works</h4><p>KeyleSSH is open source. Here’s the source code, and a demo if you’d rather just watch it run.</p><ul><li><strong>Source:</strong> <a href="https://github.com/tide-foundation/keylessh-m2">github.com/tide-foundation/keylessh-m2</a></li><li><strong>Live demo:</strong> <a href="https://demo.keylessh.com/">demo.keylessh.com</a></li><li><strong>Whitepaper:</strong> <a href="https://tide.org/whitepaper">https://tide.org/whitepaper</a></li></ul><p>We pointed Emergent Authority at a PAM on purpose, because it’s the hardest target we could think of. But the architecture applies to anything that holds authority, which is why we’re now maintainers on Red Hat’s Keycloak for identity and access, and contributing the same model to Bitwarden and other open-source projects.</p><p>The breach was always inevitable. The catastrophic damage is now optional.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=781589d2ade2" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/i-vibe-coded-a-pam-781589d2ade2">I vibe-coded a PAM</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cyber Immunity in the AI Era: Why Traditional Security Fails Agentic AI]]></title>
            <link>https://itnext.io/cyber-immunity-in-the-ai-era-why-traditional-security-fails-agentic-ai-4609290e9e92?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/4609290e9e92</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cryptography]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Wed, 29 Apr 2026 23:35:42 GMT</pubDate>
            <atom:updated>2026-05-06T08:13:38.962Z</atom:updated>
            <content:encoded><![CDATA[<p>This presentation is an adaptation of a keynote address delivered by Sasha Le, Senior Engineer, Tide Foundation at the launch event of the RMIT AWS Innovation Lab (RAIL) on 21st of April, 2026</p><h4>The Human Vulnerability</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*F3eY-nTO_NtEkr1a.jpg" /></figure><p>In 2022, a ransomware group named Lapsus$ breached some of the most sophisticated tech companies on the planet. The list included Microsoft, Nvidia, Okta, Uber, and Samsung.</p><p>The ringleader wasn’t a state-sponsored syndicate. He was a sixteen-year-old operating out of his mother’s living room. He didn’t find some obscure zero-day vulnerability, and he certainly didn’t outcode the security teams at these tech giants. He did something much simpler by bribing, blackmailing, and tricking the people inside those organizations who held the keys to the kingdom.</p><p>We have spent billions fortifying our software and systems. Yet, the people managing those systems still hold “god-like” administrative powers. For any attacker worth their salt, those people are the shortest path in.</p><p>Now, hand that attack pattern over to artificial intelligence.</p><p>We are facing superintelligent systems capable of cloning a CEO’s voice in real-time, mapping an organization’s chart in minutes, and applying perfectly timed social pressure at scale. In an era where AI can relentlessly target human psychology and software vulnerabilities, traditional cybersecurity paradigms fail at an unprecedented pace.</p><h4>The Paradox of “Vibe Coding”</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*oeo5OgmTB2Y3cuq_.jpg" /></figure><p>We cannot simply rely on AI to defend us from AI. The way we build software is changing too rapidly. We have entered the era of “vibe coding” where you simply describe what you want in plain English, and AI builds the application.</p><p>The barrier to entry has vanished, driving the global builder population from 100 million in 2023 to an estimated 1.5 billion this year. It is estimated that 85% of all production source code is now AI-generated.</p><p>While this productivity boom is incredible, it comes with a massive blind spot. We are mass-producing software with hardcoded secrets, misconfigured access, and exposed credentials at a pace no human security team can review. Vibe coding has introduced a 300% increase in vulnerabilities.</p><blockquote>Pitting one superintelligence against another to patch these millions of holes solves nothing. It’s a structural stalemate.</blockquote><p>The only way to defend against this exponentially scaling threat is to change the physics of the problem entirely. We have to remove what the AI is able to attack.</p><p>What if systems were never trusted with anything sensitive in the first place?</p><h4>Introducing Emergent Authority</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*HMaPOoh1ZtoZ77mf.jpg" /></figure><p>At Tide, we are building infrastructure based on a novel principle we call Emergent Authority.</p><p>The concept is straightforward. The actual authority to execute a sensitive action like authenticating a user or accessing critical data sits completely beyond the reach of any system or person using it.</p><p>It doesn’t live in a centralized vault (as is today’s best practice). Instead, that authority emerges just-in-time, and only when all cryptographic conditions are perfectly met.</p><p>Under this architecture, a server is effectively an oblivious proxy. It holds no power and stores no secrets. If an attacker breaches the server, they find nothing of value. If an AI socially engineers a system administrator, the breach is equally fruitless, because that administrator no longer holds direct, unilateral authority themselves.</p><p>The weakest link is rendered incapable of posing a catastrophic threat.</p><h4>Proving it with KeyleSSH</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*pO8tiq-FWcCu29z3.jpg" /></figure><p>Theoretical security is one thing, but proving it where the stakes are highest is another. We were able to test and validate this architecture through our work with the RACE Cloud Supercomputing Hub, an initiative supported by RMIT and AWS.</p><p>We applied Emergent Authority to one of the most critical and high-risk areas of IT infrastructure. That area is Privileged Access Management, known as PAM.</p><p>A PAM is the system IT teams use to manage access to the servers, firewalls, and databases running a business. By design, a traditional PAM centralizes every privileged credential an organization has. It is effectively a vault containing the keys to the kingdom. When attackers compromise these central vaults, the failure is catastrophic.</p><p>We saw this when the Lapsus$ group breached Uber by tricking an employee into handing over their credentials. This allowed the hackers to help themselves to the company’s vault. More recently, this pattern hit the U.S. federal government. A major breach at the U.S. Treasury was traced back to BeyondTrust, which is a leading PAM vendor. In that case, attackers exploited specific ‘weak spots’ in the software’s defense that allow hackers to hijack security keys. Similar vulnerabilities in vendors like Ivanti and Okta have recently impacted other agencies, including the U.S. Department of Justice. It is a stark reminder that when we centralize all our authority in one provider, we’re placing blind trust in them, and a single flaw in their code becomes a backdoor into our most sensitive institutions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*v5GS-xCXWJc-pUnR.jpg" /></figure><p>To solve this problem at its core, Tide built <a href="https://github.com/tide-foundation/keylessh-m2">KeyleSSH</a>, an open source PAM built entirely on Emergent Authority.</p><p>To understand the scale, consider that a single mid-sized utility company might have tens of thousands of servers, each with its own credentials. KeyleSSH manages the access keys to all of them but holds absolutely none of them.</p><p>When an administrator logs in through KeyleSSH, it is a Zero-Knowledge Login. No password ever leaves the device, and there is no password database on the server to be stolen. The authentication is verified cryptographically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*8bpz_ChMf4Jn_4ZF.gif" /></figure><p>When that administrator opens a session to a production server, there are no keys to store, rotate, or revoke. There is no central vault. The cryptographic authority to open the session emerges just-in-time to connect them, but neither the user nor the PAM ever possesses the key itself. Furthermore, executing commands of consequence requires cryptographic approval from other administrators, eliminating the risk of a single compromised “god-mode” account.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*w69WfsY5_veDmoeS.gif" /></figure><h4>The Infrastructure for What’s Next</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fR3goyMPacTejKR8.jpg" /></figure><blockquote>Superintelligent AI will inevitably find vulnerabilities in your software. It will inevitably social-engineer your administrators.</blockquote><p>If an architecture ensures there are no secrets stored in a vulnerable state, a successful breach yields nothing for the attacker. The goal is no longer just preventing breaches, but creating a baseline of cyber immunity. In this model, even when a network is compromised, the mathematical structure of the system renders the event structurally inconsequential.</p><p>Photographed: RMIT Deputy Vice Chancellor Research &amp; Innovation, Distinguished Professor Calum Drummond AO, Professor Alec Cameron — Vice-Chancellor and President at RMIT University, Valerie Singer — AWS General Manager of Global Education, Yuval Hertzog — Co-founder Tide Foundation, Sasha Le — Senior Engineer, Tide Foundation, Dr Ian Oppermann — ACCC Commissioner, Professor Mark Easton — Associate Deputy Vice Chancellor (Research Infrastructure), Chris Mano — Account Executive Education, AWS, Dr Robert Shen — Director of AWS Cloud Supercomputing Hub</p><p><em>Originally published at </em><a href="https://tide.org/blog/cyber-immunity-emergent-authority-ai-era"><em>https://tide.org</em></a><em> on April 29, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4609290e9e92" width="1" height="1" alt=""><hr><p><a href="https://itnext.io/cyber-immunity-in-the-ai-era-why-traditional-security-fails-agentic-ai-4609290e9e92">Cyber Immunity in the AI Era: Why Traditional Security Fails Agentic AI</a> was originally published in <a href="https://itnext.io/">ITNEXT</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[KeyleSSH: A PAM With No Single Point of Vulnerability]]></title>
            <link>https://itnext.io/keylessh-a-pam-with-no-single-point-of-vulnerability-356794a8c894?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/356794a8c894</guid>
            <category><![CDATA[ssh]]></category>
            <category><![CDATA[access-management]]></category>
            <category><![CDATA[cryptography]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 09:04:12 GMT</pubDate>
            <atom:updated>2026-08-31T07:21:50.935Z</atom:updated>
            <content:encoded><![CDATA[<h3>KeyleSSH: A New PAM Concept for IT Admins</h3><h4>For greater security and less admin overhead</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*1ylh_ay-cbsEvaFT.JPG" /></figure><p>Infrastructure security is built on a paradox: to protect your assets, you must create a catastrophic single point of vulnerability — a vault, a Certificate Authority, or a database that holds the keys to everything. The US Treasury Department <a href="https://www.bleepingcomputer.com/news/security/postgresql-flaw-exploited-as-zero-day-in-beyondtrust-breach/">learnt that lesson</a> the hard way when they put their trust in a single vendor last year.</p><p>The industry’s answer has always been “a safer box”: bastion hosts, PAM vaults, rotation policies. But a key stored anywhere is a key exposed somewhere. We call this the key-under-mat problem.</p><p>Our senior dev at Tide, Sasha, spent a few weekends trying a new PAM approach that eliminates both the key and the mat. No key left to secure or rotate, also leaves nothing to steal and zero admin overhead.</p><p>She built and open sourced the PoC, KeyleSSH, a browser-based SSH console that replaces the centralized vault with a decentralized secrets manager.</p><h3>The Architecture: Ineffable Cryptography</h3><p>KeyleSSH utilizes the <strong>Tide Cybersecurity Fabric</strong>, a network of nodes (ORKs) that operate on secrets without ever seeing them. This is based on a framework we call “Ineffable Cryptography” — secrets that are never expressed in whole, therefore can never be lost, stolen or misused.<br>Traditional Multi-Party Computation (MPC), used by advanced vaults to operate on split keys, still momentarily combines the pieces inside a Trusted Execution Environment (TEE). The protocols behind Ineffable Cryptography, however, remain entirely blind from start to end. <strong>The key is never reconstructed, not even in a TEE</strong>. The signature is mathematically constructed directly from the distributed shares.</p><p>Instead of a server holding an SSH private key and signing user requests, the signing operation itself is distributed:</p><ol><li><strong>Authentication</strong>: The user performs Zero-Knowledge authentication via OIDC (TideCloak), receiving a token bound to their device and session.</li><li><strong>Request</strong>: KeyleSSH constructs a signing request containing the raw SSH challenge bytes and human-readable metadata.</li><li><strong>Consensus</strong>: The request is sent to the decentralized Fabric. Nodes independently validate the policy (e.g., “Is User A allowed to access Server B?”).</li><li><strong>Threshold Signing</strong>: If the policy passes, nodes generate partial signatures using Tide special MPC. These are combined to form a valid Ed25519 signature.</li></ol><p>Crucially, the private key is never reassembled. Even if you compromised a node in the fabric, you would find only mathematical noise. To compromise a single key, you would need to compromise a majority of nodes simultaneously.</p><p><em>The underlying protocol has been formally analyzed over 7 years of academic research. </em><a href="https://arxiv.org/abs/2309.00915"><em>Read one of the cryptographic proofs here</em></a><em>.</em></p><h3>The Implementation: 30 Lines of Code</h3><p>Proper key management is notoriously difficult to implement. When used as part of a cryptographic protocol, like SSH, it usually requires specialized cryptography teams and complex state management. With KeyleSSH, none of that is necessary.</p><p>What makes KeyleSSH interesting is that Sasha built the core proof-of-concept in only few weekends using the <a href="https://docs.tidecloak.com/">TideCloak SDK</a>. The SDK abstracts the complex orchestration of the decentralized network and the key lifecycle management into a standard async interface.</p><p>This is the actual code that replaces the entire “secure vault” backend of a traditional PAM:</p><p><em>client/src/lib/tideSsh.ts</em></p><pre>import { IAMService } from &quot;@tidecloak/js&quot;;<br>import { TideMemory, BaseTideRequest } from &quot;heimdall-tide&quot;;<br><br>export function createTideSshSigner(): SSHSigner {<br> return async (req: SSHSignatureRequest) =&gt; {<br>   const tc = (IAMService as any)._tc;<br><br>   // 1. Pack the data (Metadata + SSH Challenge)<br>   const humanReadable = createHumanReadableInfo(req);<br>   const draft = TideMemory.CreateFromArray([humanReadable, req.data]);<br><br>   // 2. Construct the Request for the Fabric<br>   const tideRequest = new BaseTideRequest(<br>     &quot;BasicCustom&quot;, // Protocol<br>     &quot;BasicCustom&lt;1&gt;&quot;,   // Version<br>     &quot;Policy:1&quot;,         // The policy contract to execute<br>     draft,<br>     new TideMemory()<br>   );<br><br>   // 3. Attach Authorizer (The user&#39;s doken)<br>   const dokenBytes = new TextEncoder().encode(tc.doken);<br>   tideRequest.addAuthorizer(TideMemory.CreateFromArray([dokenBytes]));<br><br>   // 4. Execute Distributed Signing<br>   // The SDK handles the communication with the ORK nodes.<br>   const initialized = await tc.createTideRequest(tideRequest.encode());<br>   const sigs = await tc.executeSignRequest(initialized, true);<br><br>   return sigs[0]; // The valid Ed25519 signature.<br> };<br>}</pre><p>This snippet achieves something that previously required expensive hardware modules (HSMs) or high-risk software vaults: generating a valid signature without the signing key ever being present in memory.</p><h3>The Demise of the Rogue Admin</h3><p>Because the authority over assets is the network, away from the PAM and anyone administering it, we can enforce logic that even a root admin cannot bypass. For example, a policy can require M-of-N signatures from other admins before granting access to a production cluster.</p><p>This isn’t application-layer logic (which can be patched out by a rogue admin); it is a constraint enforced by the signing network itself. If the quorum isn’t met, the mathematical threshold for the signature is never reached.</p><h3>Current Limitations</h3><p>While the architectural model offers significant advantages over centralized PAMs, KeyleSSH as a product is currently a proof-of-concept.</p><ul><li><strong>Browser Security</strong>: The client runs in a browser. While we use Subresource Integrity (SRI), a compromised endpoint device (malware on the admin’s laptop) remains a threat vector.</li><li><strong>Host Hardening</strong>: As with any PAM, this solution solves the authentication problem. It does not replace the need for standard OS-level controls, patching, or network segmentation.</li></ul><h3>Summary</h3><p>KeyleSSH demonstrates what can be beyond the “trusted vault” era. By pushing state and authority to a decentralized fabric, we eliminate the Single Point of Failure that plagues modern infrastructure. It doesn’t eliminate every security risk, but it does eliminate the “central trust” blast radius that underpins traditional security stacks.</p><h3>Links</h3><p>Source Code: <a href="https://github.com/sashyo/keylessh/">github.com/sashyo/keylessh</a><br>Live Demo: <a href="https://demo.keylessh.com/">demo.keylessh.com</a><br>TideCloak SDK: <a href="https://docs.tidecloak.com/">docs.tidecloak.com</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=356794a8c894" width="1" height="1" alt=""><hr><p><a href="https://itnext.io/keylessh-a-pam-with-no-single-point-of-vulnerability-356794a8c894">KeyleSSH: A PAM With No Single Point of Vulnerability</a> was originally published in <a href="https://itnext.io/">ITNEXT</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond Patching: Mitigating RCE with Architectural Cryptography]]></title>
            <link>https://infosecwriteups.com/beyond-patching-mitigating-rce-with-architectural-cryptography-5b7d48797fd6?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/5b7d48797fd6</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[nodejs]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[cyrptography]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Mon, 15 Dec 2025 10:46:59 GMT</pubDate>
            <atom:updated>2025-12-15T10:46:59.719Z</atom:updated>
            <content:encoded><![CDATA[<h4>How to shut down the next “React2Shell” by removing its target</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/998/1*tKFIrhBCq6kZvdgMU86ArQ.jpeg" /></figure><h4>TL;DR</h4><p>React2Shell (CVE-2025–55182) gives attackers shell access to your server. Patching is mandatory, but if your secrets live on that server, you’ve already lost. Ineffable Cryptography removes the target asset entirely — making the RCE’s ultimate goal structurally impossible.</p><p>The disclosure of CVE-2025–55182 (React2Shell) — a CVSS 10.0 Remote Code Execution (RCE) vulnerability in React Server Components now being actively exploited and added to the CISA Known Exploited Vulnerabilities catalog — is a wake-up call for developers. With 39% of cloud environments running vulnerable React or Next.js instances, the attack surface is massive.</p><p>This vulnerability highlights a fundamental security principle: Y<strong>our system is only as secure as the most valuable secret it holds</strong>.</p><p>When an RCE occurs, an attacker gains shell access to the backend server. If that server holds the database credentials, API signing keys, or cloud secrets, the breach instantly escalates from an application exploit to a catastrophic, systemic compromise.</p><p>While robust defense-in-depth (like input sanitization and Zero Trust) is mandatory, the new cryptographic paradigm “Ineffable Cryptography” introduces a fundamental shift that allows developers to remove the target assets entirely from the attack surface.</p><h3>What Is Ineffable Cryptography and How Does It Mitigate React2Shell?</h3><p>Ineffable Cryptography is a practical application of advanced cryptographic concepts like Threshold Cryptography and secure Multi-Party Computation (sMPC). It fundamentally changes the relationship between your application and its secrets.</p><h3>1. The Core Principle: Decentralized Key Authority</h3><p>In traditional systems, a single key (a secret string, file, or token) grants “god-like” authority over data or systems. This key is stored in a single vault, file, or configuration variable on the backend server, making it the Single Point of Compromise (SPoC).</p><p>The Ineffable approach eliminates the SPoC by ensuring the full cryptographic key is never assembled or stored in one place.</p><ol><li><strong>Fragmentation</strong>: A key only exists in multiple, useless pieces (fragments).</li><li><strong>Decentralized Network</strong>: These fragments are distributed across a decentralized, independently-operated network of nodes.</li><li><strong>The Swarm Rule</strong>: To perform any action (like signing an API request or decrypting a database field), a mathematically predetermined number of fragments (a swarm) must securely cooperate. The full key is never reconstructed, even during the computation. Nodes remain forever blind to both the key and the eventual result.</li><li><strong>Quorum-Governance</strong>: Setting the parameters around the swarm-action, like who’s permitted to request, on what conditions, etc., can only be done with an explicit approval of a quorum of admins. No single admin can go rogue or risk compromise.</li></ol><h3>2. The Developer’s Benefit: Removing the RCE’s Ultimate Goal</h3><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/876bda5999ea93b679b64effb45b7fdb/href">https://medium.com/media/876bda5999ea93b679b64effb45b7fdb/href</a></iframe><p>The compromised server is reduced to a “dumb terminal” with no authority. It can only request usage of a key, but the key’s authority rests outside the server, distributed and secured by mathematics.</p><h3>Key Takeaway for All Developers</h3><p>Every developer, regardless of their role, must understand that vulnerabilities like React2Shell expose the entire application’s security surface. You should assume that other React2Shell-class weaknesses already exist or will inevitably be introduced somewhere in your software supply chain. A determined attacker will discover them before you do, which is precisely why the architectural layer of your security is just as important as the code layer.</p><ul><li><strong>Immediate Fix</strong>: Patching the insecure deserialization in the React environment is the first and non-negotiable step.</li><li><strong>Long-Term Defense</strong>: Evaluate where your most critical secrets reside. If a successful RCE on any backend server grants an attacker access to your master keys, you have a Single Point of Compromise.</li></ul><p>New cryptographic models, like the Ineffable approach represent an evolution in security architecture. Shifting from defending the bastion that holds the key (Zero Trust) to eliminating the existence of the key within any container entirely. By removing the target asset, the most severe consequences of RCE become structurally impossible.</p><p><strong>Additional reading</strong>:</p><ul><li><a href="https://medium.com/bugbountywriteup/the-god-mode-vulnerability-that-should-kill-trust-microsoft-forever-f83b8fe6e909">The God Mode Vulnerability That Should Kill “Trust Microsoft” Forever</a></li><li><a href="https://medium.com/bugbountywriteup/reimagining-cybersecurity-for-platform-developers-cbcdebcb0d74">Reimagining cybersecurity for platform developers</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5b7d48797fd6" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/beyond-patching-mitigating-rce-with-architectural-cryptography-5b7d48797fd6">Beyond Patching: Mitigating RCE with Architectural Cryptography</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to future-proof your web app’s password authentication]]></title>
            <link>https://itnext.io/how-to-future-proof-your-web-apps-password-authentication-fa385c65a75a?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/fa385c65a75a</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[passwords]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[authentication]]></category>
            <category><![CDATA[identity-management]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Tue, 30 Sep 2025 08:02:08 GMT</pubDate>
            <atom:updated>2026-08-31T06:59:44.453Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yYApSfccGArCl9Z6Wp6u5g.jpeg" /></figure><h3><strong>TL;DR</strong></h3><p>Declare:</p><pre><br>const { login } = useTideCloak();</pre><p>Add this:</p><pre>&lt;p&gt;&lt;button onClick={login}&gt;Login&lt;/button&gt;&lt;/p&gt;</pre><p>That’s it. No password storage, no hashing logic.</p><h3><strong>Some context</strong></h3><p>Traditional authentication knows dangerously too much. Its sole job is to determine whether users are who they claim to be, and nothing more. Yet today’s systems must both <strong>know how to challenge</strong> users (password/biometric) and <strong>decide</strong> if they’re valid. Whether you hash and salt passwords locally or use OAuth via a provider like Okta or Facebook, <em>someone</em> still holds secrets. When that someone is compromised, because you trusted them implicitly, you absorb the fallout.</p><p>It feels unavoidable. But what if authentication could verify a user <strong>without sending or storing secrets anywhere</strong>, and without trusting any single party to act as the authority? That’s what we’re building with TideCloak.</p><h3><strong>The problem with current approaches</strong></h3><ul><li><strong>Password hashes are targets.</strong> Modern GPU clusters can test billions of combinations per second; offline attacks thrive on stolen hashes.</li><li><strong>Social login shifts trust.</strong> You’re now betting on the <a href="https://tide.org/blog/god-mode-vulnerability-microsoft-authorityless-security">provider’s security</a> posture. Credentials (or their derivatives) still exist somewhere and can be attacked offline.</li></ul><h3><strong>How it works</strong></h3><p>Instead of storing credentials, TideCloak converts each credential into a <strong>zero-knowledge proof of a challenge</strong> that only the rightful holder can solve, then shards that proof across a network of independent nodes. No single node holds enough to reconstruct any meaningful proof, or to act as the sole authority to validate it. During login, the user engages the network to obtain a short-lived proof. The user’s credential becomes effectively <strong>ineffable</strong> — beyond the ability to be expressed anywhere, by anyone.</p><blockquote><em>INEFFABLE /ɪnˈɛfəbəl/ (adj) — too great to be expressed.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/552/1*1qnp30eFtMn3KoQZKGn4HA.jpeg" /></figure><h4><strong>Sequence (at a glance):</strong></h4><ol><li>The user clicks <strong>Login</strong>, launching a login screen with a secret session bound to the browser.</li><li>The user enters their <strong>Username</strong> and <strong>Password</strong>, performing a zero-knowledge authentication using <a href="https://link.springer.com/chapter/10.1007/978-3-031-41181-6_19">Tide’s PRISM protocol</a>, against the network.</li><li>A correctly entered credentials will manifest a cryptographic proof from the network.</li><li>Using that proof, TideCloak sends your app an OAuth2 Access Token (JWT) bound to the secret browser session.</li></ol><h4><strong>Unique properties and benefits:</strong></h4><ul><li><strong>Best of both worlds: </strong>Familiar, frictionless UX, but passkey level security.</li><li><strong>User privacy and identity ownership</strong>: Each user receives a deterministic, app-scoped<strong>,</strong> anonymized identity (untraceable and unlinkable across apps). Their master credentials are trusted to no one.</li><li><strong>Nothing to steal:</strong> no password database, no hashes, no secrets at rest — Rendering offline attacks (e.g. Rainbow) infeasible.</li><li><strong>Provable integrity:</strong> network-level and browser-level verification ensure tamper resistance and trust.</li><li>Any MITM tampering causes authentication to fail.</li><li>Resilient against token hijacking. The JWT is unusable outside its originating browser.</li></ul><h3><strong>And the implementation looks like this</strong></h3><h4><strong>Create a React app using Vite</strong></h4><pre>npm create vite@latest tidecloak-react -- --template react-ts<br>cd tidecloak-react<br>npm install @tidecloak/react</pre><h4><strong>Get TideCloak running</strong></h4><p>Run a local test instance with Docker:</p><pre>sudo docker run --name tidecloakdev \<br>  -d \<br>  -v .:/opt/keycloak/data/h2 \<br>  -p 8080:8080 \<br>  -e KC_BOOTSTRAP_ADMIN_USERNAME=admin \<br>  -e KC_BOOTSTRAP_ADMIN_PASSWORD=change-me-now \<br>  tideorg/tidecloak-dev:latest</pre><p><em>Tip: Ports 8080 (server) and 5173 (Vite) must be free.</em></p><h4><strong>Activate your TideCloak instance</strong></h4><p>Tide offers a free developer license for up to 100 users.</p><ol><li>Open: <a href="http://localhost:8080/admin/master/console/#/myrealm/identity-providers/tide/tide/settings">http://localhost:8080/admin/master/console/#/myrealm/identity-providers/tide/tide/settings</a></li><li>Sign in (Username: admin, Password: change-me-now, unless changed). You’ll land on: myrealm realm → Identity Providers → tide IdP → Settings screen.</li><li>Click <strong>Manage License</strong> → <strong>Request License</strong>.</li><li>Complete the checkout with a contact email.</li></ol><p>Within a few seconds, your TideCloak host should be licensed and activated.</p><h4><strong>Export your TideCloak adapter</strong></h4><ol><li>Go to <a href="http://localhost:8080/admin/master/console/#/myrealm/clients"><strong>Clients</strong> </a>→ mytest</li><li>Set <strong>Valid redirect URIs</strong> to http://localhost:5173/*</li><li>Set <strong>Web origins</strong> to http://localhost:5173</li><li>In <a href="http://localhost:8080/admin/master/console/#/myrealm/clients">Clients</a> menu → mytest client ID → Action dropdown → Download adaptor configs option (keep it as keycloak-oidc-keycloak-json format)</li><li>Download as tidecloak.json in your project root</li></ol><h4><strong>Edit src/main.tsx</strong></h4><p>Add imports:</p><pre>import { TideCloakContextProvider, useTideCloak, Authenticated, Unauthenticated } from &#39;@tidecloak/react&#39;<br>import tidecloakConfig from &quot;../tidecloak.json&quot;<br></pre><p>Pre-authentication view:</p><pre>function Welcome() {<br>  const { login } = useTideCloak();<br>  return (<br>    &lt;div&gt;<br>      &lt;h1&gt;Hello!&lt;/h1&gt;<br>      &lt;p&gt;Please authenticate yourself!&lt;/p&gt;<br>      &lt;p&gt;&lt;button onClick={login}&gt;Login&lt;/button&gt;&lt;/p&gt;<br>    &lt;/div&gt;<br>  );<br>}</pre><p>Post-authentication view:</p><pre>function LoggedIn() {<br>  const { logout } = useTideCloak();<br>  return (<br>    &lt;div&gt;<br>      &lt;p&gt;You&#39;re signed in!&lt;/p&gt;<br>      &lt;button onClick={logout}&gt;Logout&lt;/button&gt;<br>    &lt;/div&gt;<br>  );<br>}</pre><p>Mount the app:</p><pre>createRoot(document.getElementById(&#39;root&#39;)!).render(<br>&lt;StrictMode&gt;<br>  &lt;TideCloakContextProvider config={tidecloakConfig}&gt;<br>    &lt;Authenticated&gt;<br>      &lt;LoggedIn /&gt;<br>    &lt;/Authenticated&gt;<br>    &lt;Unauthenticated&gt;<br>      &lt;Welcome /&gt;<br>    &lt;/Unauthenticated&gt;<br>  &lt;/TideCloakContextProvider&gt;<br>&lt;/StrictMode&gt;,<br>)</pre><p><strong>Build and run</strong></p><pre>npm install<br>npm run dev</pre><h3><strong>Technical properties</strong></h3><p>TideCloak uses <strong>ineffable cryptography</strong> where authentication requires collaboration from multiple nodes, and no single node — or any subset below a threshold — can reconstruct credentials or proofs. This is a novel form of secure multi-party computation, validated through published research with RMIT and Deakin universities.</p><h4><strong>Key properties:</strong></h4><ul><li><strong>Threshold:</strong> authentication succeeds if any <strong>14 of 20</strong> nodes respond.</li><li><strong>Resilience:</strong> compromising up to <strong>14 nodes reveals nothing</strong>.</li><li><strong>Unlinkability:</strong> each service sees users as unique, unlinkable identifiers (app-scoped).</li><li><strong>Performance:</strong> on par with, or better than, major cloud IdPs; browser-bound proofs add negligible overhead.</li></ul><h3><strong>Threat model and limitations</strong></h3><h4><strong>Protects against:</strong></h4><ul><li>Database breaches revealing credentials/hashes</li><li>Insider access to user password material</li><li>Token replay in MITM scenarios (browser-bound, DPoP-style proof)</li><li>OAuth/IdP provider compromise leaking user credentials</li></ul><h4><strong>Doesn’t protect against:</strong></h4><ul><li>Keyloggers/malware on user devices</li><li>Very weak passwords</li><li>Majority network compromise above threshold (≥14/20)</li></ul><h4><strong>Current limitations:</strong></h4><ul><li>Internet connectivity required for authentication</li></ul><h3><strong>Integration simplicity</strong></h3><p>The SDK handles the cryptography. Your app calls useTideCloak() and receives a session token. Standard JWT verification works normally.</p><p>For existing systems, run traditional auth and zero-knowledge auth on TideCloak in parallel during migration, moving users over as they log in.</p><h3><strong>Open questions we’re working on</strong></h3><p>We’re in Alpha and looking for feedback on:</p><ul><li>Developer experience</li><li>Clarity of the approach</li><li>Bugs and feature requests</li></ul><p>TideCloak’s code is open source — we welcome security reviews and integration feedback.</p><h3><strong>Dive into the code</strong></h3><ul><li>One-Click Demo: <a href="https://github.com/codespaces/new/tide-foundation/tidecloak-playground?quickstart=1">https://codespace.new/tide-foundation/tidecloak-playground?quickstart=1</a></li><li>Example React project: <a href="https://github.com/tide-foundation/tidecloak-gettingstarted">https://github.com/tide-foundation/tidecloak-gettingstarted</a></li><li>SDK: <a href="https://docs.tidecloak.com/docs/Languages/React/tidecloak-react-tutorial-quickstart">https://docs.tidecloak.com/</a></li><li>Join the dev community: <a href="https://tide.org/discord">https://discord.gg/XBMd9ny2q5</a></li></ul><p>We’re building this because credentials shouldn’t be honeypots, individuals should own their identities, and organizations shouldn’t have to carry the liability of securing secrets.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fa385c65a75a" width="1" height="1" alt=""><hr><p><a href="https://itnext.io/how-to-future-proof-your-web-apps-password-authentication-fa385c65a75a">How to future-proof your web app’s password authentication</a> was originally published in <a href="https://itnext.io/">ITNEXT</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The God Mode Vulnerability That Should Kill “Trust Microsoft” Forever]]></title>
            <link>https://infosecwriteups.com/the-god-mode-vulnerability-that-should-kill-trust-microsoft-forever-f83b8fe6e909?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/f83b8fe6e909</guid>
            <category><![CDATA[identity-management]]></category>
            <category><![CDATA[authentication]]></category>
            <category><![CDATA[microsoft]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cryptography]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Fri, 19 Sep 2025 17:09:49 GMT</pubDate>
            <atom:updated>2025-10-24T02:14:40.041Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*AjegA8x4wfGJI_LjYCy6aw.jpeg" /></figure><p><strong>How One Token Could Have Compromised Every Microsoft Entra ID Tenant on Earth, And Why It’s Time for Authorityless Security</strong></p><p>Recently, security researcher Dirk-Jan Mollema disclosed CVE-2025–55241, a vulnerability so catastrophic that it reads like fiction: a single token, obtained from any test tenant, could have granted complete administrative control over every Microsoft Entra ID (Azure AD) tenant in the world. Every. Single. One.</p><p>Let that sink in. Dirk-jan uncovered a path where a lab account was all it took to open the doors to every Microsoft cloud customer, from Fortune 500 corporations to the smallest startups.</p><p>This isn’t just another vulnerability. It’s another smoking gun that proves our entire approach to cybersecurity is fundamentally broken.</p><h4><strong>Another God Mode Vulnerability in a Long Line of “Never Should Have Happened” Breaches</strong></h4><p><strong>The short version for non-technical readers:</strong> An attacker could have become anyone in any Microsoft cloud tenant, accessed anything, and done it all completely undetected with what appeared to be legitimate access that bypassed all security controls. Think of it as finding a master key that opens every door in every building in the world, while also making you invisible to all security cameras.</p><p><strong>The technical details:</strong> Microsoft’s backend uses something called “Actor tokens” for service-to-service communication. Due to a critical validation flaw, these tokens could be used across tenant boundaries, granting attackers the ability to read all user data, group memberships, application permissions, even BitLocker keys, completely undetected. When ready to strike, they could create new Global Admin accounts or take over existing identities. <a href="https://dirkjanm.io/obtaining-global-admin-in-every-entra-id-tenant-with-actor-tokens/">Read the full technical breakdown here</a>.</p><p>The vulnerability existed because Microsoft’s architecture requires services to have god-like powers to function. They built a system where ultimate authority had to be trusted to someone, and that trust became the vulnerability.</p><h4><strong>“If We Can’t Trust Microsoft, Who Can We Trust?”</strong></h4><p>For decades, enterprises have justified their security decisions with a simple question: “If we can’t trust Microsoft, who can we trust?” It’s been the ultimate defense when things go wrong: at least you chose the industry standard.</p><p>This vulnerability, along with countless others across the industry, demolishes that logic. Microsoft joins a growing list of “trusted” vendors that suffered catastrophic breaches: <a href="https://www.bleepingcomputer.com/news/security/okta-october-data-breach-affects-all-customer-support-system-users/">Okta’s exposure of all customer support system users</a>, <a href="https://www.bleepingcomputer.com/news/security/cisco-warns-of-cslu-backdoor-admin-account-used-in-attacks/">Cisco’s hidden backdoor account exploited in attacks</a>, <a href="https://www.bleepingcomputer.com/news/security/beyondtrust-says-hackers-breached-remote-support-saas-instances/">BeyondTrust’s remote support SaaS compromise</a>, and <a href="https://cyberpress.org/severe-vulnerabilities-ztna-solutions/">severe zero-trust flaws in Check Point, Zscaler, and Netskope</a>. Each had the resources, expertise, and certifications. Each failed catastrophically. It’s inevitable.</p><p>What should really concern you is the invisibility of these flaws. Behind all the security promises and compliance badges, customers had no way of knowing these Actor tokens existed or carried such sweeping powers. You can’t audit what you can’t see, and you can’t secure what you must blindly trust.</p><p>The uncomfortable reality is that this is just the vulnerability that was discovered and disclosed responsibly. How many similar flaws are lurking in these vast codebases? How many have already been found by nation-state actors who are quietly exploiting them? When you have systems this complex, with this much centralized authority, vulnerabilities aren’t anomalies. They’re mathematical certainties.</p><h4><strong>The Ultimate Supply Chain Attack Vector</strong></h4><p>Large platforms like Microsoft aren’t just targets, they’re supply chain force multipliers. Compromise Microsoft, and you compromise everyone. This vulnerability proved that definitively.</p><p>Consider the exponential nature of the attack path discovered: starting from a single compromised tenant, an attacker could jump to every tenant that had guest users, then to every tenant those tenants had connections to. Within minutes, the majority of all Entra ID tenants worldwide could have been mapped and compromised.</p><p>Microsoft’s own tenant would likely be one of the first compromised (since Microsoft consultants are guests in countless customer tenants), and from there, every major service provider would be one step away. The entire global business infrastructure, falling like dominoes.</p><p>This isn’t a bug. It’s the natural consequence of building systems where ultimate authority must be blindly trusted to the system, and those that administer it.</p><h4><strong>Why “Authority” Is the Real Vulnerability</strong></h4><p>First, let’s define what we mean by “authority” in the digital world: it’s the power to grant access, approve actions, or enforce rules within a system. It’s what lets someone or something open the door to sensitive data, change configurations, or allow operations to run. Think of it as the master key to your digital kingdom.</p><p>The root cause of this Microsoft vulnerability wasn’t poor coding or lack of testing. It also isn’t correct to say that it’s the need to trust Microsoft. It’s more accurately what we’re trusting Microsoft with — Authority.</p><p>As long as someone or something holds it, it can be exploited. Today’s systems are built with someone, somewhere, having the ability to do anything. Maybe it’s an administrator, maybe it’s the vendor, maybe it’s a service account, maybe it’s a sub-system, but that ultimate power exists. And it doesn’t need to. Modern cryptographic techniques have made it possible to build systems where no single entity needs or has complete authority, where authority itself is distributed across multiple independent parties who must collaborate to act.</p><p>Traditional security, even “Zero Trust” security, doesn’t eliminate authority. It just moves it around. You stop trusting the network perimeter, but now you must blindly trust your Identity Provider. You implement defense in depth, but at the bottom of that depth sits an all-powerful system that, if compromised, renders every other defense meaningless.</p><p>The Actor token vulnerability perfectly illustrates this. It doesn’t matter how many security policies you had, how much you spent on monitoring, or how well-trained your security team was. When the Identity Provider itself is compromised at the most fundamental level, all of that becomes theater.</p><h4><strong>The Path to Authorityless Security</strong></h4><p>We need systems where inevitable breaches don’t have catastrophic consequences. We need architectures where authority doesn’t exist in any single place that can be compromised.</p><p>Imagine a world where:</p><ul><li>No single entity, not even the platform vendor, can access your sensitive data</li><li>No administrator, no matter how privileged, can override the system without oversight</li><li>No supply chain compromise can grant universal access</li><li>No vulnerability, no matter how critical, can provide god-mode to attackers</li></ul><p>This isn’t just better for security, it’s for a better ecosystem. Platform vendors can demonstrate they’re beyond reproach. Businesses can verify their security rather than hope for it. Users can trust systems because they don’t have to. Mistakes will be made, breaches will continue, but platforms will be cyber-immune to the consequences.</p><p>The technology to eliminate centralized authority already exists. Forward-thinking vendors are adopting it because they recognize the core architectural flaw, don’t want the liability, want to build fast and innovate without losing sleep over breaches, want to simplify compliance, and want customers to trust them because they no longer have to.</p><p>Consider how this alternative approach works: authority over user identities, administrative actions, and business data is distributed in the form a key who’s pieces live across a decentralized network. No single node, organization, or administrator ever holds complete power. When someone needs access to data:</p><ol><li>Their identity is verified through distributed authentication where no single node knows their credentials</li><li>Authorization requires cryptographic consensus from multiple independent servers</li><li>Cryptographic keys needed to decrypt data or sign access tokens never exist in any one place, only their cryptographic actions manifest through the collaboration of multiple independent nodes</li><li>Even if an attacker compromises the IAM or multiple nodes holding key fragments, they cannot forge access or decrypt data</li></ol><p>The crucial difference: there is no Actor token. There is no god mode. There is no single point of authority that, if compromised, hands over the kingdom.</p><p>In such a distributed system, the keys that could elevate privileges or forge tokens remain out of anyone’s reach. So even with a vulnerability similar to Microsoft’s, attackers still couldn’t access customer data, impersonate users, or grant themselves permissions. Mathematical guarantees ensure that authority remains distributed, even in the face of compromise.</p><h4><strong>The Uncomfortable Questions We Must Ask</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*B2-Z0LEB1qcAHc5cp_iI7w.jpeg" /></figure><p>Next week, another critical vulnerability will be discovered. Another emergency patch will be rushed out. Another post-mortem will be written. But the fundamental problem will remain: we’re trusting single entities with absolute authority over our most critical systems.</p><p>Consider the numbers: despite over $300 billion spent on cybersecurity annually, breaches caused over $10 trillion in damage last year. Things are only getting worse. This isn’t a failure of execution, it’s a failure of architecture.</p><p>How many undiscovered Actor token-equivalent vulnerabilities exist right now? How many have been found by adversaries who are silently exploiting them? How long can we pretend that “trust Microsoft” or “trust Google” or “trust any single vendor” is an acceptable security strategy?</p><p>The answer is clear: we can’t. Not anymore.</p><h4><strong>The Future Is Authorityless</strong></h4><p>The Actor token vulnerability isn’t just another security incident. It’s more proof that our entire approach to security is fundamentally flawed. As long as ultimate authority exists in any single place, that place will be compromised.</p><p>The future belongs to systems where authority doesn’t exist in any single place. Where cryptographic proofs replace blind trust. Where mathematical guarantees supersede vendor promises. Where the inevitable breach doesn’t mean inevitable catastrophe.</p><p>This isn’t about replacing one vendor with another. It’s about eliminating the need for vendors to hold any authority at all, replacing it with distributed, verifiable, mathematically-guaranteed security.</p><p>Microsoft has fixed this particular vulnerability. But the architecture that made it possible, the existence of centralized, god-like authority, remains. And it will remain until we collectively decide that trusting any single entity with everything is no longer acceptable.</p><p>The question isn’t whether another Actor token-level vulnerability exists. It’s when it will be discovered, and by whom.</p><p>Isn’t it time we stopped gambling with that question?</p><p><em>The authorityless architecture described in this piece is not theoretical. Advances in distributed cryptography now make it practical, economical, and verifiable at scale. At Tide Foundation, our peer reviewed research in ineffable cryptography and the development of platforms like </em><a href="https://docs.tidecloak.com/"><em>TideCloak</em></a><em> show how identity and access can be managed without trusting ultimate authority to any vendor, administrator, or insider. The goal is not for customers to “trust Tide,” but to remove the need for trust in any single entity at all.</em></p><p><em>Learn more at </em><a href="https://tide.org/"><em>tide.org</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f83b8fe6e909" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/the-god-mode-vulnerability-that-should-kill-trust-microsoft-forever-f83b8fe6e909">The God Mode Vulnerability That Should Kill “Trust Microsoft” Forever</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ghost in the Network]]></title>
            <link>https://infosecwriteups.com/ghost-in-the-network-51b73b17fe50?source=rss-bba93caba371------2</link>
            <guid isPermaLink="false">https://medium.com/p/51b73b17fe50</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[identity]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[privacy]]></category>
            <category><![CDATA[authentication]]></category>
            <dc:creator><![CDATA[Tide Foundation]]></dc:creator>
            <pubDate>Tue, 10 Jun 2025 08:01:38 GMT</pubDate>
            <atom:updated>2025-06-11T01:49:53.886Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*a6rdlN4tX2TibH_Up0vnCA.jpeg" /></figure><p><strong>A car engine is only supposed to start with the right key. Yet, as every heist movie shows, pry off the steering column, twist a couple of wires, and the car is yours! The rule that only the key can bring a machine to life turns out to be optional.</strong></p><p>Corporate networks operate on the same wishful thinking. Firewalls, cloud apps, and databases are meant to stay locked until a user shows the right credentials, but attackers slip in through overlooked bugs, compromised vendors, or sloppy settings — just to name a few. The login screen feels like a steel vault door. In practice, it’s drywall.</p><p>Breaches are now a daily headline. One day it’s a payroll platform, the next a hospital system. The details change, but not the script. A target system already holds the power it needs to function, so once an intruder finds an unintended path to the controls, they can “start the engine”. Unlike a car thief who can only work one dashboard at a time, a hacker can hot-wire a thousand organizations, simultaneously. A single coding slip can — and does — trigger catastrophe.</p><p>We keep piling hurdles on real users, like one-time codes, CAPTCHA riddles, fancy passkeys — while the software beneath never truly needs the user’s key. It’s an architectural flaw we paper over with an ever-growing constellation of security tools. Even with all these tools at our disposal, IBM’s 2024 Cost of a Data Breach report pegs the average breach at $4.88 million, a 10 percent jump in a single year.</p><p>The time has come to fundamentally rethink this approach. What if we could build systems where the user truly is the missing key?</p><p>Imagine a car whose engine could only work with its specific driver present. The driver themselves becomes a missing puzzle piece required for this particular engine to start. Without the driver, this car is just an expensive hunk of steel that not even the manufacturer can operate. Hot-wiring the car would be futile because only the driver can complete the circuit.</p><p>Recent advances in <a href="https://tide.org/blog/rethinking-cybersecurity-for-developers">zero-knowledge cryptography</a> make this vision achievable for digital systems. We can now treat a user like that missing puzzle piece — one that forever sits outside the system itself. When the real user shows up, the missing piece powers only the slice of data they’re entitled to and vanishes when they leave. Beyond that interaction, the email platform, CRM, or AI assistant remains a harmless pile of ones and zeros.</p><p>This approach flips the entire threat model. An attacker must first become the user, and even then, the damage stops at that single account. Administrators can’t impersonate customers, malware can’t rummage through decrypted records, and compliance shifts from promises to mathematical certainty.</p><p>Security teams will still patch bugs and monitor logs, but that work becomes maintenance rather than crisis response. Software vendors that embrace this “user-as-key” architecture will become significantly harder targets, and insurers, regulators, and customers will take notice.</p><p>The best safety upgrades seem obvious in hindsight. Platforms that require the rightful user’s presence to function will not only shed breach anxiety and liability; they cultivate the kind of trust that turns customers into advocates.</p><p>Soon, any service that leaves sensitive data accessible without an authorized user present will seem as outdated as a car without seatbelts.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=51b73b17fe50" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/ghost-in-the-network-51b73b17fe50">Ghost in the Network</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>