<?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 Donica Briggs on Medium]]></title>
        <description><![CDATA[Stories by Donica Briggs on Medium]]></description>
        <link>https://medium.com/@donica.briggs?source=rss-83faffdbe566------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*odTYaEycXMt1VarPfcVb4Q@2x.jpeg</url>
            <title>Stories by Donica Briggs on Medium</title>
            <link>https://medium.com/@donica.briggs?source=rss-83faffdbe566------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 03:22:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@donica.briggs/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[Insider Threats: When Trust Is the Vulnerability]]></title>
            <link>https://medium.com/@donica.briggs/insider-threats-when-trust-is-the-vulnerability-ca63b07fe195?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/ca63b07fe195</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[insider-threat]]></category>
            <category><![CDATA[devsecops]]></category>
            <category><![CDATA[cloud-security]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Mon, 05 Oct 2026 17:41:16 GMT</pubDate>
            <atom:updated>2026-10-05T17:41:16.432Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*xtFTow4YnIbGX1dS9xnEEA.jpeg" /></figure><p>Most security conversations default to defending a perimeter, keeping the wrong people out. Insider threats flip that entirely, the risk isn’t someone breaking in, it’s someone who already has a login, credentials and an organization’s trust. This makes insider threats easy to underestimate.</p><h4>Tesla — 2018: The Insider with Root Access</h4><p>In June 2018, Tesla CEO Elon Musk told employees the company had been the victim of “<em>extensive and damaging sabotage</em>.” According to Tesla’s own lawsuit and reporting from CSO Online and Dark Reading, a process technician named Martin Tripp made direct, unauthorized code changes to Tesla’s Manufacturing Operating System using false usernames, exporting gigabytes of sensitive data, including video of Tesla’s manufacturing systems, to outside parties. Tripp’s actions were reportedly the result of being passed over for a promotion.</p><p>This case often gets cited in insider threat training not because of the sabotage itself, but due to how small the gap was between an employee with legitimate access and an employee causing extensive damage to production systems. No external firewall to breach. No credentials to steal. Just someone with an existing, trusted account and a motive.</p><h4>Twitter — 2020: When the Insiders Are the Attack Surface</h4><p>Not every insider incident starts with a malicious employee, sometimes the employee is the victim and the attacker’s real target is the trust placed in an employee’s access. In July 2020, attackers compromised 130 high-profile Twitter accounts including former President Barack Obama, Elon Musk and Apple’s corporate account to push a Bitcoin scam that netted over $118,000 in a matter of hours.</p><p>According to Twitter and a detailed timeline, the attackers didn’t exploit a software vulnerability. They ran a phone spear-phishing campaign against Twitter employees, impersonating IT staff to extract credentials, first from a lower-level employee without admin access, then using that foothold to social-engineer additional employees who had access to Twitter’s internal account management tools. A 17-year-old was later identified as the mastermind.</p><p>Insider threats don’t require a disgruntled employee, just a normal, well-meaning one who gets fooled. Twitter’s own post-mortem pointed to gaps in real-time monitoring of internal tool usage as part of what enabled attackers to operate for hours before being shut down.</p><h4>Different Insiders, One Shared Root Cause</h4><p>One a malicious employee acting on a personal grievance, the other, well-meaning employees manipulated by outside attackers. The Tesla and Twitter incidents both trace back to the same structural issue, access that outpaced oversight. Martin Tripp’s account access allowed him to touch production systems directly, under false usernames, without triggering any alerts. Twitter’s admin tools let a handful of employees take over any account on the platform, lacking real-time monitoring to catch the unusual usage.</p><p>Insider risk isn’t about whether an employee is trustworthy. It’s about whether the access they hold is scoped to what their role requires, and whether anyone’s watching for unusual use of that access.</p><h4>How the Cloud Raises the Stakes</h4><p>On-premises, an insider’s reach is often limited due to physical access and network segmentation. In a cloud environment, a single set of credentials with console access can touch production infrastructure, customer data or billing globally, instantly from anywhere.</p><p>A few cloud-specific risk factors:</p><p><strong>-Cloud consoles concentrate power that used to be distributed</strong>. A cloud admin with broad IAM permissions can spin up resources, exfiltrate data to external storage, or tear down infrastructure from a single browser session; no physical datacenter access required, no second system to compromise.</p><p><strong>-Shared and standing credentials in cloud environments.</strong> API keys, service account credentials and long-lived access tokens often get shared across teams or embedded in automation.</p><p><strong>-SaaS admin tools.</strong> An organization running customer data through a SaaS platform is trusting that platform’s employees the same way Twitter’s users were trusting Twitter’s. When evaluating a cloud vendor, internal access controls become part of the risk surface.</p><p><strong>-Cloud logging makes detection possible. </strong>Services like CloudTrail, Azure Monitor and Google Cloud’s Audit Logs can generate log trails to catch anomalous internal activity such as an account suddenly accessing resources outside its normal pattern, unusual API call volume or access from an unexpected region. The gap is usually in whether data is actively being monitored versus just retained for an after-the-fact investigation.</p><h4>Insider Threat Prevention</h4><p>Reducing the risk of insider threats safeguards data, prevents financial losses and preserves organizational trust. Tooling such as Privileged Access Management (PAM), paired with just-in-time access that grants elevated permissions for a defined period of time, limits the unreviewed, persistent access that both Martin Tripp and Twitter’s compromised accounts leveraged to enact lasting damage.</p><p><strong>-Least-privilege access enforced continuously</strong>. Access should map to current role(s) and needs, not accumulate over time as employees change positions or take on side projects.</p><p><strong>-Behavioral monitoring, not just perimeter monitoring</strong>. Tools that catch external attackers rarely catch insiders, because insiders are, by definition, already inside. Detecting unusual internal activity, such as code changes under an unfamiliar username, requires monitoring built specifically for that blind spot.</p><p><strong>-Attention to human signals, without resorting to a surveillance culture</strong>. A demotion or a resignation notice don’t guarantee bad behaviour; it should be considered when conducting access reviews instead of ignored out of discomfort.</p><p><strong>-Segregation of duties for high-impact systems</strong>. No single account should be able to make unreviewed changes to something as consequential as a manufacturing operating system or a platform’s account management tooling.</p><p><strong>-Cloud-native controls</strong>. Cloud Access Security Brokers (CASBs) and User and Entity Behavior Analytics (UEBA) platforms exist to catch anomalous internal activity that perimeter tools miss, flagging a service account for pulling unusual data volumes, or an admin console login from a pattern that doesn’t match the user’s history.</p><h4>Trust But Verify</h4><p>In incidents previously covered, the focus was defending the perimeter against external threats. Insider threats are a reminder of risks that can exist inside a perimeter amongst employees who already have credentials. The risk of an insider threat doesn’t shrink in the cloud, if anything, the reach a single compromised or malicious account has increases. Trust isn’t a vulnerability in and of itself. Trust without ongoing verification is.</p><p>Sources: Tesla v. Tripp court filings; CSO Online; Dark Reading; Fortra, Verizon Data Breach Investigations Report; Twitter (<em>official incident disclosure</em>).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ca63b07fe195" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Trust is a Contract: The Importance of SLAs]]></title>
            <link>https://medium.com/@donica.briggs/trust-is-a-contract-the-importance-of-slas-b8dbce4f4b66?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/b8dbce4f4b66</guid>
            <category><![CDATA[sla]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[vendor-risk-management]]></category>
            <category><![CDATA[service-level-agreement]]></category>
            <category><![CDATA[cloud-security]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sun, 27 Sep 2026 15:52:44 GMT</pubDate>
            <atom:updated>2026-09-29T13:20:08.366Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/554/1*NGXfLeMvLQKsoWVXG9nmLg.jpeg" /></figure><p>A Service Level Agreement (SLA) is a contract that defines what a vendor owes, such as uptime guarantees, response times, remediation commitments and what happens when they fail to deliver. In cloud environments, when trusting a third party with infrastructure, data or security tooling not directly controlled, an SLA is the formal leverage when something goes wrong.</p><h4>When the Security Vendor Is the Outage</h4><p>On July 19, 2024, CrowdStrike a cybersecurity company whose Falcon platform is deployed specifically to protect endpoints, pushed a faulty configuration update that caused a logic error in its Windows sensor. According to CrowdStrike’s own account and CISA’s advisory, the incident resulted in widespread system crashes and blue screens across machines running the Falcon sensor. This wasn’t a cyberattack. It was, by CrowdStrike’s own admission, a defect in a routine update.</p><p>The scale was staggering. Estimates from Bitsight put the number of affected devices at approximately 8.5 million, disrupting airlines, hospitals and financial institutions worldwide. Delta, American and United were forced to ground all flights and reporting from the Cloud Security Alliance (CSA) list the Fortune 500 impact at an estimated $5.4 billion.</p><h4>Contractual Liability</h4><p>CrowdStrike wasn’t hacked. It didn’t get breached. Its own software broke itself. Vendor risk doesn’t care whether the vendor’s mistake was hostile or accidental. A strong SLA won’t prevent an outage, however, it forces a conversation about acceptable risk before being locked into a vendor relationship, and it determines if an organization has any recourse when, not if, something goes wrong.</p><p>A security vendor’s own product caused one of the largest IT outages in history, and for many affected organizations, the financial and operational damage exceeded whatever contractual remedy their SLA provided. As aviation analysts writing for Cirium pointed out in the aftermath, this is a scenario that should nudge organizations toward more stringent SLAs, ones where a vendor shares in the financial impact of an outage they caused.</p><h4>SLAs &amp; the Cloud</h4><p>Cloud computing didn’t invent the SLA, but it raised the stakes on it. Under the shared responsibility model that governs various major Cloud Service Providers (CSPs), the provider secures the underlying infrastructure, but organizations are responsible for what is built on top of it, accountable to customers if it fails. An SLA is the document that defines where the provider’s responsibility ends, in dollar terms, when something goes wrong.</p><p>There’s also a scale problem unique to cloud environments. A single vendor failure doesn’t just take down one server, it can simultaneously take down every customer relying on that region or service. Major cloud providers openly acknowledge this trade-off in their SLA terms. For example, AWS’s service commitments typically guarantee monthly uptime percentages and offer service credits, calculated as a percentage of the bill, when thresholds aren’t met.</p><p>Multi-tenancy compounds this further. When infrastructure runs on a shared cloud platform, an outage or misconfiguration affecting the provider’s broader environment can impact consumers even if their configuration is flawless. This is why redundancy across regions, availability zones, and vendors has become standard in cloud architecture.</p><h4>The Fine Print</h4><p>CrowdStrike is a reminder that trust in a vendor, even a security vendor, isn’t a substitute for a contract. Too often organizations treat SLAs as something legal reviews, revisiting it only when there’s a crisis. There are several key principles worth noting:</p><p>·<strong>Liability proportional to the impact.</strong> A service credit doesn’t come close to covering the cost of a multi-day operational outage. Terms should scale with business impact.</p><p>·<strong>Redundancy for critical functions.</strong> A vendor should not be a single point of failure for mission-critical operations. Multi-vendor redundancy should be leveraged when practical.</p><p>·<strong>Revisiting of SLAs as dependency increases.</strong> The agreement signed when a vendor was a minor tool shouldn’t be the same one governing once they’re deeply embedded in critical operations. Terms should be revisited as the relationship’s importance changes.</p><p>·<strong>Clear incident communication.</strong> Beyond uptime, an SLA should specify how fast and how transparent a vendor communicates during an active incident. The difference between a contained crisis and a chaotic one comes down to information, not just technology.</p><p>Thanks for joining me!</p><p>Next up, we’ll do a deep dive on insider threats and how easy they are to underestimate.</p><p>Sources: CrowdStrike incident report; CISA advisory; Bitsight; Cloud Security Alliance; Mass General Brigham public statements; Cirium aviation analysis; AWS service level agreement documentation.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b8dbce4f4b66" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Inside the Norsk Hydro Ransomware Attack]]></title>
            <link>https://medium.com/@donica.briggs/inside-the-norsk-hydro-ransomware-attack-e0ca105ffbd2?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/e0ca105ffbd2</guid>
            <category><![CDATA[ransomware]]></category>
            <category><![CDATA[critical-infrastructure]]></category>
            <category><![CDATA[nist-framework]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 19 Sep 2026 20:26:51 GMT</pubDate>
            <atom:updated>2026-09-19T20:26:51.297Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/678/1*DItNwdG_-eimygf_T9anXA.jpeg" /></figure><p>Each incident of this critical infrastructure cybersecurity series so far has involved a misconfiguration. This one is another case study in criminal ransomware, and it’s also a good example of how to respond when prevention has failed.</p><p>On the evening of March 18, 2019, Norsk Hydro, one of the world’s largest aluminum producers, with roughly 35,000 employees across 40 countries discovered its IT systems were under active ransomware attack. According to CyberScoop’s reporting, the attack started in the company’s U.S.-based plants and by the next morning, the company’s CFO confirmed publicly that production had been suspended at multiple facilities.</p><p>Hydro’s employees across every business area likely felt the impact. According to a Cybersecurity and Infrastructure Security research case study, the ransomware, known as LockerGoga locked files across 23,000 computers and 3,000 servers.</p><h4>How It Started</h4><p>Reporting from Microsoft’s own security blog, and based on Hydro’s internal forensic investigation, the infection can be traced three months prior, to December 2018, not March. Attackers weaponized an email attachment sent from a trusted customer’s employee to a Hydro employee, as part of what looked like a legitimate ongoing business conversation. That attachment carried a trojan.</p><p>Hydro’s antivirus caught the trojan a few days after the initial attack. By then, forensic investigators later determined, it was already too late. The attackers had a foothold. According to analysis from Recorded Future’s Insikt Group, attackers spent months moving through Hydro’s network largely undetected, using compromised credentials and Active Directory to position themselves for a company-wide deployment.</p><p>When LockerGoga finally activated on March 18, it wasn’t a smash-and-grab. It was a culmination of months of patient and quiet positioning.</p><h4>A Defining Decision</h4><p>According to Microsoft’s account of the incident, company executives made several fast, consequential decisions in the first few hours of the crisis. They refused to pay the ransom, they proactively and publicly disclosed the attack rather than trying to contain the narrative, and they made the call to switch industrial control systems across numerous plants into manual mode rather than risk further spread.</p><p>The manual mode decision resulted in tens of thousands of employees that normally used automated systems having to complete tasks by hand. A disruptive and expensive choice, but it contained the damage to IT systems rather than letting uncertainty about OT exposure force a full shutdown, echoing the same IT/OT tension that shaped Colonial Pipeline’s response years later. The financial impact reached an estimated $71 million, according to Microsoft’s reporting, but Hydro avoided the prolonged, undefined outage that hits organizations who don’t have a clear containment plan.</p><p>Many companies treat a ransomware attack as an event to control, not share. Hydro did the opposite. They held public press conferences within a day, publishing ongoing updates and eventually releasing detailed internal accounts of what happened and how they responded. That transparency became a case study that’s now taught across the security industry. It’s an example of a victim organization treating public disclosure as part of the response rather than a liability to be managed.</p><h4>Preventative Cloud Solutions</h4><p>The three-month dwell time between initial infection and full activation is what a modern cloud-native security stack is designed to shrink. Cloud-based email security and attachment sandboxing, inspecting attachments in an isolated environment before they ever reach an inbox, could have caught the weaponized attachment. Trusted-sender relationships, like the legitimate customer conversation this attack rode in on, are the blind spot behavioral sandboxing is designed to catch.</p><p>Once the trojan was inside, cloud-based identity protection and privileged access management would have narrowed what the attackers could do with the compromised credentials and Active Directory access. Cloud SIEM and UEBA platforms, tuned to flag unusual authentication patterns and lateral movement across three months instead of three hours, are built precisely to surface the kind of slow-burn positioning that let this attack reach a company-wide scale.</p><p>On the response side, cloud-based, immutable backup and disaster recovery is what could have kept the production network ransomware isolated from reach.</p><p>A cloud-isolated backup strategy is a modern lever that makes refusing to pay a ransom a less costly decision. Native cloud solutions don’t guarantee an attack will never happen, it means it will get caught within hours or days instead of months.</p><h4>The Recurring Lesson</h4><p>The Norsk Hydro breach wasn’t the result of an overlooked security control. The phishing email could have happened anywhere, to anyone. What differentiates this incident is what happened undetected during the months of quiet lateral movement and then critically, what happened in the first 24 hours of the ransom.</p><p>Prevention is important, and each article in this series has presented that case. Norsk Hydro is a reminder that response planning matters just as much. Knowing in advance whether you’ll pay a ransom and having a plan for switching critical operations to manual mode are decisions that should be made in minutes during a real incident, instead of for the first time under pressure.</p><p>This closes out the critical infrastructure series. A pipeline, a water plant, a healthcare clearinghouse and now an aluminum producer. Four different sectors, four different root causes, but the same recurring lesson. Technology that fails and how well an organization survives it comes down to the groundwork laid before an incident ever occurs.</p><p>Thanks for joining me!</p><p>Next up, we’ll cover Service Level Agreements (SLAs), the contract that defines what vendors are responsible for delivering.</p><p>Sources: CyberScoop; Microsoft Security Blog; Recorded Future/Insikt Group; CISA/CyOTE Case Study Analysis; Moody’s IncidentAnalysis.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e0ca105ffbd2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Attack That Broke U.S. Healthcare Billing]]></title>
            <link>https://medium.com/@donica.briggs/the-attack-that-broke-u-s-healthcare-billing-c6ae8324e9aa?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/c6ae8324e9aa</guid>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[nist-framework]]></category>
            <category><![CDATA[critical-infrastructure]]></category>
            <category><![CDATA[ransomware]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 12 Sep 2026 13:40:36 GMT</pubDate>
            <atom:updated>2026-09-12T14:25:26.628Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gdkeC533NwA-FNFKv3p-lQ.png" /></figure><p>The last critical infrastructure attack we covered was on a water plant servicing 15,000 people. This time, we’ll dissect the missing security setting on a remote access portal that disrupted medical claims processing for the nation’s largest healthcare system.</p><h4>The Largest Healthcare Data Breach in U.S. History.</h4><p>Change Healthcare, a UnitedHealth Group subsidiary, processes a substantial share of medical claims and electronic prescriptions across the U.S. According to TechTarget, roughly a third of all American medical claims flow through its systems. On February 21, 2024, the company detected a ransomware attack that took its systems offline for weeks, disrupting prescription fulfillment, insurance billing, and payment processing at hospitals, pharmacies and physician practices nationwide.</p><p>The ransomware group ALPHV, also known as BlackCat, claimed responsibility. According to UnitedHealth Group CEO Andrew Witty’s later congressional testimony, the attackers got in nine days before deploying ransomware, using compromised credentials to log into a Citrix remote access portal.</p><h4>The Root Cause</h4><p>The Citrix portal did not have MFA enabled. Not a novel exploit or a zero-day. Just compromised credentials likely obtained through a prior breach or credential-stealing malware, walked straight through a remote access point with no second verification step.</p><p>Once inside, the attackers spent days moving laterally through Change Healthcare’s network before triggering the ransomware payload. Staged attacks like this one often generate detectable signals such as unusual authentication patterns and abnormal data access volumes. Behavioral monitoring tuned to catch this activity could have interrupted the attack before the encryption ever occurred.</p><h4>The Fallout</h4><p>UnitedHealth Group’s disclosures put the financial impact at over $872 million, and the company advanced more than $9 billion to healthcare providers who couldn’t process payments during the outage. According to HIPAA Journal and Breachsense’s incident tracking, Optum the UnitedHealth subsidiary handling the response, paid a $22 million ransom to ALPHV.</p><p>Paying the ransom didn’t guarantee the data stayed contained. After the payout, the ransomware group’s leadership pulled an exit scam, stealing the payment from their own affiliate. The unpaid affiliate kept a copy of the stolen data and shopped it to RansomHub, a secondary ransomware operation, demanding an additional payment.</p><h4>A Single Point of Failure</h4><p>The Change Healthcare incident isn’t just a ransomware story, it’s systemic risk. A huge share of an entire sector’s payment infrastructure ran through one company, and one missing MFA configuration at that company became the single point of failure for hospitals, pharmacies and patients who had no direct relationship with Change Healthcare and no way to route around the outage. Millions of patients found themselves paying out of pocket because the billing systems weren’t able to process claims.</p><p>There seems to be a pattern for critical infrastructures. Whether or not it’s disclosed anywhere in a business continuity plan, when a single vendor becomes infrastructure for an entire sector, that vendor’s security posture becomes everyone’s security posture.</p><h4>Native Cloud Solutions</h4><p>A cloud identity platform would have enforced no-exception MFA on every remote access portal, including third-party and legacy systems like the Citrix gateway. Conditional access policies that evaluate device health, location and behavior before granting access can also provide an additional layer of defense that a stolen password alone can’t satisfy.</p><p>During the nine-day dwell time, cloud-based UEBA and SIEM tooling would have flagged unusual internal authentication and data-access patterns that preceded the ransomware deployment, ideally early enough to interrupt the attack before encryption.</p><p>At the architectural level, this incident is a case for reducing vendor concentration risk. Cloud-based multi-vendor redundancy for critical functions like claims processing means one compromised provider doesn’t become a vulnerability for an entire sector.</p><h4>Back to Basics</h4><p>The Change Healthcare attack impacted over 192 million people and is deemed the largest healthcare data breach in U.S. history. Nearly every major critical infrastructure incident we’ve covered traces back to a basic security control that existed and was well understood, it simply wasn’t enforced. Enforce MFA on every remote access point, with no exceptions for legacy systems.</p><p>Thanks for joining me!</p><p>Next up, we’ll cover the ransomware attack on one of the world’s largest aluminium producers, impacting 35,000 employees and 40 countries.</p><p>Sources: TechTarget; HIPAA Journal; Cloudskope incident analysis; Breachsense.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c6ae8324e9aa" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Oldsmar Water Plant Hack]]></title>
            <link>https://medium.com/@donica.briggs/the-oldsmar-water-plant-hack-367e21052275?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/367e21052275</guid>
            <category><![CDATA[ics-security]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[critical-infrastructure]]></category>
            <category><![CDATA[cloud-security]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Tue, 08 Sep 2026 18:36:58 GMT</pubDate>
            <atom:updated>2026-09-08T18:36:58.595Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nrw6D4MBg7nehwuox_pnpQ.png" /></figure><p>Previously we covered how a leaked password shut down the fuel supply to half of the East Coast. The Oldsmar incident is even more frightening. It was an attempt to poison a town’s drinking water.</p><p>On February 5, 2021, an operator at the water treatment plant in Oldsmar, Florida, a city of approximately 15,000 people, noticed their cursor moving on its own. According to reporting from CNN and Cybersecurity Dive, it happened once in the morning and was dismissed as a supervisor remotely troubleshooting. Around 1:30 that afternoon, it happened again, except this time, whoever was in control of the mouse navigated to the software managing water chemistry increasing the sodium hydroxide level (<em>lye</em>) from 100 parts per million to 11,100 parts per million. That’s a hundredfold increase of a chemical typically used in small doses to control water acidity and corrosion.</p><p>The plant operator watching the screen caught it within minutes and reverted the change before the highly concentrated, hazardous substance reached the water supply. No one was harmed, but the fact it got as far as it did is the story.</p><h3>A Public Scare</h3><p>According to a CISA advisory issued in the aftermath and Ars Technica’s reporting, the Oldsmar plant’s systems were running Windows 7, an operating system Microsoft had stopped supporting over a year earlier, with no firewall between the plant’s computers and the open internet. Remote access ran through TeamViewer, a collaborative software platform, where multiple employees shared the same login credentials, so there was no way to distinguish a legitimate supervisor logging in from anyone else who’d obtained that same password.</p><p>As cybersecurity firm Dragos’s CEO Rob Lee told CNN, the outdated operating system wasn’t the core issue. The attacker didn’t need to exploit a software vulnerability. The remote access software was exposed to the internet in a way that let anyone who had the credentials (<em>guessed or purchased</em>) log in.</p><h3>A Complicated Footnote</h3><p>Years later, reporting from CyberScoop raised several questions about whether this was an external hack or a local employee’s error being misread as an intrusion. The FBI never issued a final public ruling. I’m including this nuance because a “<em>definitely happened this exact way</em>” story is more satisfying than an unresolved one, but the unresolved version seems to be closer to the truth.</p><p>Even if the “<em>hacker</em>” turns out to have been an internal misunderstanding, the security posture that was in place would have made a genuine external attack just as problematic. Shared passwords, no MFA, no firewall, unsupported software and remote industrial control access exposed directly to the internet. Regardless if this specific incident was a true external intrusion, the configuration was a vulnerability existing in a water system that served 15,000 people.</p><h3>Cloud Solutions That Could Have Helped</h3><p>The Oldsmar hack is a case where cloud-based tooling could have proved useful. Cloud solutions are ideal when dedicated security can’t be staffed. A cloud-hosted Zero Trust Network Access (ZTNA) platform, replacing direct-to-internet tools like TeamViewer would have required per-user authentication and MFA before anyone reached the control interface, it would have removed the plant’s systems from open internet exposure entirely. Cloud-based identity management with unique, centrally issued credentials would have replaced the shared-password problem outright.</p><h3>Convenience vs Security</h3><p>There is a theme that seems to recur across critical infrastructures, small municipal utilities with tight budgets, running on aging systems and remote access tools adopted for convenience rather than security. This is not a criticism of the people, it’s merely the identification of a resourcing and prioritization problem, which are gaps cloud-based, centrally managed security tooling is well-positioned to close.</p><p>The Oldsmar incident is less a story about needing cutting-edge technology and more about operators needing affordable, centrally managed alternatives to consumer-grade remote access tools. Unique credentials per user, MFA on remote access points and no direct internet exposure for control systems isn’t advanced guidance, it’s the floor.</p><p>Thanks for joining me!</p><p>Next up, a ransomware attack on a healthcare claims processor that disrupted a third of the entire U.S. medical billing system, all due to a missing MFA checkbox.</p><p>Sources: CNN; ABC News; Cybersecurity Dive; CyberScoop; CISA; Ars Technica.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=367e21052275" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Colonial Pipeline Shut Down]]></title>
            <link>https://medium.com/@donica.briggs/the-colonial-pipeline-shut-down-9e06d6485c25?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e06d6485c25</guid>
            <category><![CDATA[nist-rmf]]></category>
            <category><![CDATA[critical-infrastructure]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[ransomware]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 29 Aug 2026 19:01:23 GMT</pubDate>
            <atom:updated>2026-08-30T07:21:56.574Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/738/1*COE6ZX5c_YFjf0CfK00keA.jpeg" /></figure><p>Now that we’ve wrapped up our compliance series, let’s pivot to one of my favorite topics: cybersecurity of critical infrastructures. Water, energy, healthcare, power grids are all systems where a cybersecurity failure doesn’t just mean a data breach. It can result in hospital diversions, power outages or a city without clean water.</p><p>First up, the attack that turned a reused password into a national emergency.</p><h4>Ransomware-as-a-Service</h4><p>On May 7, 2021 Colonial Pipeline which, per GovTech’s reporting, moves over 100 million gallons of fuel a day from Texas to New Jersey, shut down its operations after discovering a ransomware infection. Ransomware is malicious software used by attackers to hold data and devices hostage, requiring victims to remit a form of payment to regain access.</p><p>The shutdown, done as a precaution while the company attempted to figure out how far the security compromise had spread, triggered fuel shortages and panic buying across the Southeastern U.S. Headlines around the world captured images of gas stations surrounded by long lines of vehicles. The incident was so serious that President Biden declared a state of emergency.</p><p>Ransomware-as-a-service (RaaS) is a criminal, pre-built ransomware kit rented or sold to hackers. According to The Hacker News’ coverage of the incident, using a RaaS model, DarkSide the group behind it, took a cut from affiliates that carried out the attacks.</p><h4>The Entry Point</h4><p>During his testimony before a House committee, Mandiant’s Charles Carmakal stated the attackers gained access through a single VPN account. There was no phishing email, no zero-day exploit and no sophisticated malware delivery chain. According to Carmakal’s testimony, an employee reused a password on a site that was breached elsewhere, that same password worked on an old inactive VPN account that despite being unused, was still active and lacked multi-factor authentication (MFA).</p><p>The attackers logged in and from there, per Carmakal’s account, they moved laterally across Colonial’s IT network, stealing close to 100 gigabytes of data in approximately two hours, deploying ransomware. Confirmed by Colonial’s CEO during Senate testimony, Colonial paid a $4.4 million ransom in bitcoin, a payment later partially recovered by the Department of Justice.</p><h4>IT vs OT Impact</h4><p>The ransomware incident impacted Colonial’s business information technology (IT) systems such as billing, scheduling and corporate email, not the operational technology (OT) that physically controls the pipeline. Colonial shut down pipeline operations because they weren’t confident the infection hadn’t reached the OT side, and because their billing systems couldn’t track fuel deliveries or payments reliably enough to continue operating safely.</p><p>Although attackers never accessed the pipeline control systems, the outcome was still a multi-day shutdown of critical fuel infrastructure. In critical infrastructure environments, IT and OT don’t have to share a direct compromise for an IT incident to force an OT shutdown. The uncertainty itself is the disruption.</p><h4>Threat Vectors</h4><p>The root cause of the Colonial attack can be mapped to several vectors: lack of MFA on remote access accounts, active unused accounts and credential reuse despite a known breach data. None of this required cutting-edge tooling. Implementation of NIST 800–53 Risk Management Framework (RMF) security controls such as <em>Access Control (AC)</em>, <em>Identification and Authentication (IA</em>), <em>System and Communications Protection (SC)</em> and <em>Contingency Planning (CP)</em> could have prevented or mitigated the ransomware incident.</p><h4>Cloud-Native Solutions</h4><p>Every gap in this story has a cloud-native answer that’s mature and widely available today. A cloud identity provider with enforced MFA across all remote access, with no exceptions for old or inactive accounts would have blocked the login outright even with the reused password. Automated identity governance tools can flag and disable stale accounts on a schedule instead of relying on someone remembering they exist.</p><p>A cloud-based SIEM tuned to flag logins from unusual locations or after long dormancy, could have also caught the anomaly in real-time. Credential-exposure monitoring services, which scan breach dumps for organizational credentials, would have flagged the reused password as compromised before anyone ever tried to use it against Colonial’s VPN.</p><h4>Closing the Gap</h4><p>Colonial Pipeline is a case I reference whenever someone argues critical infrastructure attacks require nation-state sophistication. The gap is usually consistency, not capability. The attack that took down the fuel supply for half the East Coast only needed a leaked password and an inactive account somebody forgot to disable.</p><p>Cloud-hosted or otherwise, if you’re responsible for securing a critical infrastructure the lesson isn’t to just watch out for advanced adversaries. Common measures such as auditing dormant accounts and enforcing MFA anywhere remote access exists are essential for a secure and resilient infrastructure.</p><p>Thanks for joining me!</p><p>Next up, a water treatment plant, a shared TeamViewer password and a hacker who attempted to poison a town’s drinking water.</p><p>Sources: The Hacker News; Huntress Threat Library; GovTech; Forbes Technology Council.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e06d6485c25" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Shadow IT in The Cloud]]></title>
            <link>https://medium.com/@donica.briggs/shadow-it-in-the-cloud-e5bec94f3bb3?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/e5bec94f3bb3</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[shadow-it-governance]]></category>
            <category><![CDATA[cloud-security]]></category>
            <category><![CDATA[compliance]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 22 Aug 2026 18:38:23 GMT</pubDate>
            <atom:updated>2026-08-23T08:15:58.426Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/738/1*LgUCIA3s4o87wyhREu5XlQ.jpeg" /></figure><p>Somewhere in an organization right now, a <em>“temporary”</em> AWS Virtual Private Cloud (VPC) someone spun up months ago still has a production database attached to it. It’s not a secret and noone’s hiding it maliciously. It’s just, there. Running quietly outside the security perimeter. Welcome to Shadow IT, the problem that doesn’t announce itself with an alert, it just shows up in an incident report.</p><p>Shadow IT is any technology such as applications, cloud services, devices or scripts in use by an organization without the knowledge, approval, or visibility of the IT and cybersecurity teams responsible for protecting it. Note what’s absent: malice. The person that spun up the rogue VPC instance doesn’t see themselves as a threat actor.</p><h3>How It Happens</h3><p>Most people don’t just wake up and decide to build an unauthorized cloud environment out of spite. It happens gradually, and it’s often for a reasonable reason:</p><p>·<strong>Frustration with Approved Tools.</strong> A sanctioned platform can be slow or require multiple approvals. The free alternative simply works. Guess which one will get used?</p><p>·<strong>Extensive Procurement Lead Time.</strong> A team needs a tool now for a deadline, and they don’t want to wait for a full cyber assessment. So, someone procures it and decides they’ll loop in cyber later. Unfortunately, later rarely comes.</p><p>·<strong>Noone Thought to Ask</strong>. Much of shadow IT isn’t a rebellious workaround, it’s someone genuinely not realizing that spinning up a cloud storage bucket or a new SaaS trial is something cyber needs to know about.</p><p>·<strong>The Org Chart Doesn’t Match the Tech Stack</strong>. A department with its own budget and its own priorities doesn’t always see itself as needing IT or cyber’s blessing for a tool that only affects their workflow.</p><h3>What It Looks Like in the Cloud</h3><p>For some, shadow IT invokes an image of a rogue laptop or an unauthorized USB drive. In a cloud environment, it’s far more distributed and easier to do without anyone noticing.</p><p>Here’s a few common examples:</p><p>·<strong>Shadow SaaS </strong>— A team procures a project management tool, a design platform or an application without cyber oversight or explicit approval. It goes into production immediately, which is why it doesn’t get flagged.</p><p>·<strong>Shadow Storage </strong>— Personal Google Drive, Dropbox or iCloud accounts storing company files because they’re familiar and already installed. Convenient for the person using it, but invisible to those responsible for knowing where sensitive data is stored.</p><p>·<strong>Shadow Infrastructure — </strong>A developer spins up an EC2 instance, a storage bucket or a proof-of-concept environment in a personal or loosely governed cloud account to move faster on a prototype. It works, it ships and then it just exists, quietly running in production long after anyone remembers it was supposed to be temporary.</p><p>·<strong>Shadow AI </strong>— The newest and fastest-growing category, where users paste proprietary code, client data or internal documents into a public AI chatbot to get a quick answer. No malicious intent, but also no visibility into where that data goes once it’s submitted, or whether it’s being used to train a model outside of organizational control.</p><p>·<strong>Shadow Integrations </strong>— Someone connects a sanctioned tool to an unsanctioned one through an API key or an OAuth grant to automate a workflow. This creates a data pipeline that cyber never approved and, in many cases, has no visibility of.</p><p>Each of these are collectively invisible and cumulatively problematic. Multiply this across hundreds of employees over several years, and most organizations are running a second, undocumented cloud environment stacked on top of the one cyber actually manages.</p><h3>The Blast Radius</h3><p>Shadow IT isn’t new, it existed back when the worst-case scenario was an unauthorized USB drive or a personal laptop. What’s changed is the blast radius. Spinning up a fully functional cloud environment can take merely minutes; no server closet and no IT department noticing a new workstation humming under someone’s desk.</p><p>An unauthorized system isn’t just a laptop anymore, it’s a database, an API or a storage bucket with a public-facing endpoint, sitting outside of security controls a cyber team spent months or even years implementing. You can have flawless IAM policies, a robust Zero Trust architecture and a current SBOM for every system you know about. However, none of it matters for the system a cyber team doesn’t know exists.</p><h3>You Can’t Secure What You Can’t See</h3><p>The truth about cloud security is that despite every framework we’ve talked about in this series, hardening, access control, continuous monitoring and incident reporting, it’s only successful based on the assumption you know what you’re protecting. Shadow IT disrupts this assumption at the root, quietly without ever tripping an alarm; there’s no alert configured for a system cyber doesn’t know to monitor.</p><h3>Banning Is Not an Option</h3><p>The instinctive response is to lock everything down, block every unsanctioned tool and threaten disciplinary action. However, people will only just get better at not disclosing their workaround.</p><p>Here’s a few recommendations to close the gap:</p><p>-Implement a fast, streamlined procurement process.</p><p>-Conduct regular, discovery scans across cloud environments to validate what’s running.</p><p>-Research and suggest a usable set of sanctioned tools, so unauthorized options aren’t pursued as easy or convenient alternatives.</p><p>The technical controls from this series can only work against the systems your cyber team knows exist. Shadow IT is a reminder that your actual attack surface and your documented attack surface are two different things. The gap between them will likely be where the next incident quietly occurs.</p><p>Thanks for joining me and stay tuned for the next article!</p><p>Up next, we’ll discuss cloud security for critical infrastructures.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e5bec94f3bb3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CIRCIA: “We’ll Report It Eventually” Is No Longer Acceptable]]></title>
            <link>https://medium.com/@donica.briggs/circia-well-report-it-eventually-is-no-longer-acceptable-9ce17b7bbd6c?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/9ce17b7bbd6c</guid>
            <category><![CDATA[compliance]]></category>
            <category><![CDATA[circia]]></category>
            <category><![CDATA[cisa]]></category>
            <category><![CDATA[critical-infrastructure]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 15 Aug 2026 17:47:38 GMT</pubDate>
            <atom:updated>2026-08-15T21:17:46.715Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/774/1*ALDDgiw-MqKrywItPVufrA.jpeg" /></figure><p>Back in my SBOM article, I mentioned CIRCIA and the shift toward mandatory, quick-turn incident reporting. If an ATO is about confirming a system is secure enough to run, and STIGs are about validating it’s hardened, CIRCIA is a different topic entirely. It’s what happens after something goes wrong, and how fast the right people find out.</p><h3>The Convenience Gap</h3><p>Instead of whenever it’s convenient or legally unavoidable, the Cyber Incident Reporting for Critical Infrastructure Act (CIRCIA) requires certain critical infrastructure organizations to report significant cyber incidents and ransomware payments to the federal government within defined timeframes.</p><p>The “<em>whenever it’s convenient</em>” statement isn’t sarcasm. In my upcoming critical infrastructure series, we’ll do a deep dive into several critical infrastructure security incidents such as the Colonial Pipeline, Change Healthcare and Ukraine power grid attacks. A recurring theme across major incidents, cyber or otherwise is the gap between “<em>we know something happened”</em> and “<em>the appropriate people outside our organization know something happened.</em>” CIRCIA exists to close this gap.</p><h3>Why CIRCIA Exists</h3><p>In the past, incident reporting in the U.S. consisted of patchwork. Some sectors had reporting obligations tied to specific regulators, others didn’t. Many organizations reported breaches only when a state breach notification law forced their hand, or when the story broke publicly and reporting became a public relations necessity rather than a proactive choice.</p><p>The problem with patchwork is that intelligence about active threats is more useful in real-time than it is postmortem weeks or months later. If a ransomware group is actively exploiting a vulnerability against pipeline operators or water utilities, the government’s ability to warn and posture the next potential victim depends entirely on how quickly the incident becomes known. A breach disclosed months after the fact, once litigation and PR strategy have had their say, does very little to protect the next organization in an attacker’s queue.</p><p>CIRCIA, signed into law in 2022, was created to change the timeline specifically for critical infrastructure sectors, sectors where a delayed response doesn’t just mean lost data, it can mean lost power, lost water treatment or lost access to medical care.</p><h3>Incident Requirements</h3><p>Critical infrastructure sectors are required to report two categories of events:</p><p>· <strong>Significant cyber incidents.</strong> These must be reported within a tight window. The law specifies 72 hours from the point an organization reasonably believes an incident has occurred.</p><p>· <strong>Ransomware payments.</strong> These have a more aggressive time constraint. The law requires 24 hours from when a payment is made, regardless of whether the incident has been fully reported yet.</p><p>Reporting must be submitted to Cybersecurity and Infrastructure Security Agency (CISA), which uses aggregated data to spot patterns across all sectors, issuing warnings and providing support to the victim organization.</p><h3>The Timeline</h3><p>During an incident, instinctively there’s a desire for more time, more certainty about the scope, more clarity on a root cause and more control over the narrative. CIRCIA is built to resist this instinct. A 72-hour window forces reporting to happen while an incident is still being actively investigated, not after the story has been fully assembled and sanitized.</p><p>This will be uncomfortable for organizations that are used to controlling their own disclosure timeline. However, that’s exactly the point. Threat intelligence that arrives after a threat has already moved to its next target isn’t threat intelligence, it’s history.</p><h3>Two Different Clocks</h3><p>Meeting the 72-hour or 24-hour mandate assumes an organization can actually detect and characterize an incident that fast. This assumption is where the real work is. It’s where everything connects. An accurate SBOM helps a team answer “<em>are we affected” </em>within hours. Likewise, continuous monitoring, built as a habit rather than a pre-audit scramble, is what identifies an incident on day one.</p><p>None of CIRCIA’s reporting requirements are achievable without the groundwork: inventory, monitoring and a solid incident response process that doesn’t need to be built from scratch the moment something breaks.</p><h3>Beyond the Regulated Sector</h3><p>Even outside of formal CIRCIA obligations, the underlying shift is worth internalizing. The expectation for how fast an organization should know what’s happening inside their system continues to be compressed. “<em>We’ll figure it out</em>” is no longer an acceptable posture during an incident.</p><p>Whether it’s an ATO, STIGs, Zero Trust principles, SBOMs or CIRCIA, the pattern is the same. Know what your system is running, harden it, control who has access, monitor it for anomalies and be prepared to validate and respond to incidents.</p><p>Thanks for joining me and stay tuned for the next article!</p><p>Next on the docket is shadow IT. We’ll discuss how something as simple as a forgotten, temporary AWS account attached to a production database can be problematic.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9ce17b7bbd6c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[SBOMs: What’s Inside Your Software Is A Big Deal]]></title>
            <link>https://medium.com/@donica.briggs/sboms-whats-inside-your-software-is-a-big-deal-42a9e9566a2f?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/42a9e9566a2f</guid>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[devsecops]]></category>
            <category><![CDATA[software-security]]></category>
            <category><![CDATA[sbom]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 08 Aug 2026 17:04:54 GMT</pubDate>
            <atom:updated>2026-08-08T17:04:54.330Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B7Q5obi0w7ubVZl80EpT2g.png" /></figure><p>Last time, we broke down why Zero Trust isn’t something you just buy off a shelf. This time, I want to cover something a little less conceptual and a lot more literal, knowing what’s inside the software you’re running.</p><p>That’s what a Software Bill of Materials (SBOM) is for, and if you haven’t come across the term yet, I’m sure you will soon. In many software supply chains, it’s evolved from a niche compliance detail to a genuine requirement.</p><h3>What It Is</h3><p>An SBOM is a complete, structured inventory of every component that makes up a piece of software, down to the individual libraries and dependencies. Think of it like an ingredients list. It’s not just this application performs x, y, z, but a literal list, such as the specific library and the specific version used by a specific dependency.</p><h3>Why This Matters</h3><p>Here’s an uncomfortable truth about modern software, almost nothing is built entirely from scratch anymore. A typical application might consist of predominately open-source and third-party components glued together with a relatively small amount of code a team wrote themselves.</p><p>The problem is that when a vulnerability gets disclosed in an open-source library, most organizations have no reliable way to answer a simple question, “<em>is that library running anywhere in our system?</em>” This question can take days or weeks to answer, if at all. Cyber teams will scramble, pinging their engineering team hoping someone remembered what dependencies they’d pulled in months ago. Unfortunately, attackers can move a lot faster than a scramble.</p><p>A handful of major supply chain incidents have made this gap impossible to ignore. When a widely-used component turns out to have a serious vulnerability, organizations that have an accurate SBOM can answer “<em>are we affected</em>” within minutes. The ones that don’t will be attempting to figure out what they are running while the vulnerability is already being exploited.</p><h3>What’s Inside</h3><p>For every component, a proper SBOM typically includes:</p><p>· <strong>What it is</strong> — The specific library, package or module.</p><p>· <strong>What version</strong> — The exact version number in use.</p><p>· <strong>Where it came from</strong> — The supplier or source of the component.</p><p>· <strong>What it depends on</strong> — The dependencies, which have their own dependencies.</p><p>· <strong>Licensing information </strong>— The terms that govern the use of the component.</p><p>Multiply that across every dependency, and the dependency-of-a-dependency chains, and a single application can have an SBOM listing hundreds or even thousands of components.</p><h3>Why This Isn’t Just a Cyber Issue</h3><p>SBOMs exist at an interesting intersection. Cyber teams need them to respond quickly to vulnerabilities, legal and compliance teams want them to manage license risk, and increasingly regulators and customers require them as a condition of doing business at all.</p><p>You can’t quickly assess a vulnerability or an incident if you don’t already know what’s running in your environment. SBOMs are increasingly becoming a standard part of software procurement; you can’t manage a supply chain risk you can’t see.</p><h3>Where It Gets Hard</h3><p>Generating an SBOM once isn’t the hard part, there are various tools that can scan a codebase and create one. The challenge is maintaining its accuracy. Software can change; new dependencies get added, versions get bumped and older components remain hidden when teams forget to remove them.</p><p>An SBOM that’s accurate on the day it’s generated and stale a month later isn’t much better than not having one, it just creates false confidence. The organizations getting real value out of this are the ones treating SBOM generation as a continuous part of their build pipeline, not a one-time compliance artifact.</p><h3>Beyond Compliance</h3><p>Even outside of a regulatory mandate, the underlying discipline is to know what’s running in your environment, down to the dependency level, and be able to answer “<em>are we affected</em>” the moment a vulnerability gets disclosed.</p><p>Thanks for joining me and stay tuned for the next article!</p><p>Next time we’ll pivot to something a little different: CIRCIA and what the shift toward mandatory incident reporting means for teams on the ground.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=42a9e9566a2f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Zero Trust: Product vs Principle]]></title>
            <link>https://medium.com/@donica.briggs/zero-trust-product-vs-principle-50eff7ede324?source=rss-83faffdbe566------2</link>
            <guid isPermaLink="false">https://medium.com/p/50eff7ede324</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[cloud-security]]></category>
            <category><![CDATA[devsecops]]></category>
            <category><![CDATA[zero-trust]]></category>
            <dc:creator><![CDATA[Donica Briggs]]></dc:creator>
            <pubDate>Sat, 01 Aug 2026 15:41:38 GMT</pubDate>
            <atom:updated>2026-08-01T16:08:03.168Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/637/1*DSjsCSBASeYDO2jirHeOqA.png" /></figure><p>In my last article we covered STIGs, the detailed checklists that turn “<em>harden your systems</em>” from vague advice into something specific and checkable. This time, let’s tackle a term that gets thrown around loosely: Zero Trust.</p><p>If you’ve recently sat through a vendor pitch, you’ve probably heard some version of “<em>our platform delivers Zero Trust</em>.” But Zero Trust isn’t something you just purchase. Zero Trust means no user, device or system is trusted by default, regardless of whether it’s inside or outside the network perimeter; every request must be verified, each time. That’s it. It’s a principle, not a product.</p><h3>The Castle Model</h3><p>Traditional network security is designed a lot like a castle, built with a strong perimeter such as firewalls, VPNs, and network segmentation. Once inside that perimeter users are generally trusted; a device internal to the network gets more benefit of the doubt than an external one. A user who’s already authenticated once often stays trusted for the rest of their session, across everything they access.</p><p>This model made a lot of sense when “<em>inside the network</em>” meant something specific, like a physical office or a defined set of on-premises servers. It makes a lot less sense now. Cloud infrastructure doesn’t have a clean perimeter. Once an attacker gets past the perimeter through a phished credential, a compromised device or a vulnerable third-party connection, the castle model gives them full trusted access to everything inside. Zero Trust exists because this type of single point of failure continues to get exploited.</p><h3>What It Looks Like in Practice</h3><p>Instead of trust based on network location, Zero Trust evaluates every request against several factors continuously, not just upon initial login:</p><p>· <strong>Who is making the request: </strong>Verified identity, not just a session token that’s assumed to still be valid.</p><p>· <strong>What device they’re using: </strong>Is it a known, compliant device, or something unmanaged and unpatched?</p><p>· <strong>What they’re attempting to access: </strong>Least-privilege access to the specific resource needed, not broad access to everything on the network.</p><p>· <strong>What are the conditions: </strong>Unusual location, unusual time, unusual behavior pattern, all of which can trigger additional verification or denial, even for an otherwise valid user.</p><p>None of these individually are new ideas. What’s different is treating them as a unified, ongoing verification process instead of a one-time gate at the perimeter.</p><h3>Why “Buy This Product” Is the Wrong Framing</h3><p>Here’s where the marketing can be misleading. A single product might help implement a portion of the Zero Trust approach: enabling better identity verification, better network segmentation, better device compliance checking. However, no single product delivers Zero Trust on its own because Zero Trust is an architecture designed by how multiple systems work together, not a feature you toggle on.</p><p>There are 7 core Zero Trust security principles in the defense sector:</p><p>· <strong>User</strong>: Continuously authenticate, authorize and monitor user identities and privileges (identity governance, MFA, PKI, least privilege).</p><p>· <strong>Device</strong>: Verify and secure every device connection to the network, including real-time posture assessment, patching and endpoint detection/response.</p><p>· <strong>Network &amp; Environment</strong>: Segment and isolate resources (micro-segmentation) so lateral movement is limited even after a breach.</p><p>· <strong>Application &amp; Workload</strong>: Secure applications and workloads at the code, container and runtime level.</p><p>· <strong>Data</strong>: Categorize, tag and encrypt data with access controls tied to data sensitivity instead of network location.</p><p>· <strong>Visibility &amp; Analytics</strong>: Aggregation across all pillars to detect anomalies and drive automated risk decisions (SIEM/SOAR).</p><p>· <strong>Automation &amp; Orchestration</strong>: Automated policy-driven responses to enforce security decisions at machine speed rather than relying on manual intervention.</p><p>Teams often purchase tools marketed as a Zero Trust solution and consider the job done, while access requests remain broadly trusted the moment a user pasts a single login. That’s not Zero Trust. That’s the old castle model repackaged.</p><h3>The Reality</h3><p>Implementing Zero Trust is challenging, and it’s not just a technology problem. It requires knowing exactly what your resources are, who legitimately needs access to what, and the willingness to enforce least-privilege access even when it’s inconvenient for people who are used to broad access. That’s an organizational and cultural shift just as much as a technical one, and it’s usually the part that stalls Zero Trust initiatives.</p><h3>Beyond the Buzzword</h3><p>Whether or not an organization ever formally adopts a Zero Trust architecture as a named initiative, the underlying shift is worth internalizing. Don’t design access around where a request is coming from, design it around continuously verifying who’s asking and what they’re asking for. Every time, not just once. That’s a much harder standard to meet than a strong perimeter. It’s also a much harder one to fully compromise with a single mistake.</p><p>Thanks for joining me and stay tuned for the next article!</p><p>Up next, we’ll shift gears slightly and dig into SBOMs and why knowing what’s inside your software is important.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=50eff7ede324" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>