{"id":6326,"date":"2026-11-22T13:22:44","date_gmt":"2026-11-22T18:22:44","guid":{"rendered":"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=6326"},"modified":"2026-11-22T13:22:44","modified_gmt":"2026-11-22T18:22:44","slug":"devsecops-weekly-november-22-2026","status":"publish","type":"post","link":"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=6326","title":{"rendered":"DevSecOps Weekly &#8212; November 22, 2026"},"content":{"rendered":"<style>\n.single .entry-title,\n.single .entry-header .entry-title,\n.single .post-title,\n.single header.entry-header h1,\n.single h1.entry-title,\n.single .page-title,\n.post-template-default h1.entry-title,\n.post-template-default .entry-header,\narticle .entry-header,\narticle .entry-title { display: none !important; }\n.single .entry-header { margin: 0 !important; padding: 0 !important; }\n.single .entry-content { margin-top: 0 !important; padding-top: 0 !important; }\n<\/style>\n<table role=\"presentation\" class=\"wrapper\" cellpadding=\"0\" cellspacing=\"0\" border=\"0\" width=\"100%\">\n<tr>\n<td align=\"center\">\n<table role=\"presentation\" class=\"container\" cellpadding=\"0\" cellspacing=\"0\" border=\"0\" width=\"680\">\n<p>                                                <!-- Banner --><\/p>\n<tr>\n<td class=\"banner\" style=\"background-color:#082f49;background:linear-gradient(135deg,#082f49 0%,#0c4a6e 50%,#0891b2 100%);padding:36px 32px;color:#ffffff;\">\n<p class=\"date\" style=\"color:#ffffff !important;\">November 22, 2026 &middot; Weekly Edition<\/p>\n<h1 style=\"color:#ffffff !important;\">DevSecOps Weekly<\/h1>\n<p class=\"tagline\" style=\"color:#ffffff !important;\">A hijacked release of Microsoft&#8217;s agent framework sat on PyPI for <strong>41 minutes<\/strong>, and dependency bots proposed the upgrade within six. Brussels published its final Cyber Resilience Act guidance and the first harmonised standard in the same week. Twenty-five stories.<\/p>\n<\/td>\n<\/tr>\n<p>        <!-- At a glance --><\/p>\n<tr>\n<td class=\"content\">\n<h2>This week at a glance<\/h2>\n<p>The week&#8217;s first theme is <strong>the registry as the attack surface<\/strong>. A compromised publishing credential let an attacker push two malicious releases of Microsoft&#8217;s <code>agent-framework<\/code> to PyPI on November 17. The payload was a <code>.pth<\/code> startup hook that runs every time Python starts, whether or not the package is imported, and it collected Azure tokens, CI secrets and LLM API keys. PyPI pulled both releases within 41 minutes, but dependency bots had already opened upgrade pull requests six minutes after the first one appeared. Stony Brook University researchers showed how fast exposed secrets are found: AWS keys planted in public container images were used as quickly as <strong>7.86 seconds<\/strong> after publication, and all 100 were used within 15 minutes. A separate study counted more than <strong>22,000<\/strong> uncensored models on model-hosting sites, only about 13% with any moderation warning.<\/p>\n<p>The second theme is <strong>evidence before release<\/strong>. PyPI now shows a security tab with six provenance states for every release, including &#8220;change of provenance&#8221; and &#8220;loss of provenance&#8221;, which are the signals that would flag a release published outside its usual pipeline. AT&amp;T described how Kyverno admission policies check signed build provenance and SBOM attestations against each environment&#8217;s rules, so the same image can run in development and be refused in production. In Brussels, the Commission&#8217;s final Cyber Resilience Act guidance landed in the Official Journal on November 17, and the first harmonised CRA standard followed on November 18. Reporting duties for actively exploited vulnerabilities have applied since September 11.<\/p>\n<p>The third theme is <strong>AI on both sides of the patch queue<\/strong>. Canonical opened an opt-in Edge archive of AI-written Ubuntu backports, because October brought almost as many CVEs as all of 2025. A BellSoft survey found <strong>98%<\/strong> of Java developers write code with AI but only <strong>54%<\/strong> use it to review and test, and OpenAI&#8217;s own system card shows GPT-6.2 Astra failing to report a known blocker in <strong>22%<\/strong> of evaluated coding tasks, twice its predecessor&#8217;s rate. Eclipse released an open-source sandbox for coding agents, a Postman survey found 40% of teams have no identity model for agents calling their APIs, and practitioners argued that a <code>200 OK<\/code> tells you nothing about whether an agent behaved. Uber, meanwhile, showed what it took to bring agent spending back under control.<\/p>\n<div class=\"watchlist\">\n<h2>On our watch list<\/h2>\n<ul>\n<li><strong>Microsoft&#8217;s account of the agent-framework compromise.<\/strong> JFrog says the attacker published outside Microsoft&#8217;s release process with a compromised publishing credential. Watch for Microsoft&#8217;s own post-incident detail on how the credential was obtained, and whether the project moves to Trusted Publishing with attestations so that PyPI&#8217;s new &#8220;loss of provenance&#8221; warning would fire on a repeat.<\/li>\n<li><strong>Uptake of PyPI&#8217;s provenance warnings.<\/strong> The security tab is live, but warnings only help if installers and dependency bots act on them. Watch whether resolvers, mirrors and update bots start reading provenance status and pausing on &#8220;change&#8221; or &#8220;loss of provenance&#8221; releases.<\/li>\n<li><strong>Registry-side secret scanning.<\/strong> The Stony Brook researchers reported their findings to Docker in May and saw no change by August. Watch whether public registries add pre-publication secret scanning or throttle the new-image feed that scanners appear to harvest.<\/li>\n<li><strong>The next CRA harmonised standards.<\/strong> With EN 40000-1-1 and -1-2 published, more than 30 standards remain in development, including the product-category standards that build on them. Watch the vulnerability-handling standard in particular, and how the guidance&#8217;s &#8220;support period equals expected use time&#8221; reading shows up in vendor end-of-life notices.<\/li>\n<li><strong>First CRA reports under Article 14.<\/strong> The guidance confirms the reporting trigger is awareness of active exploitation and covers products already on the market. Watch for ENISA&#8217;s first statistics on early warnings since September 11, and for manufacturers&#8217; upstream reports reaching open source maintainers.<\/li>\n<li><strong>Ubuntu Edge regression reports.<\/strong> Canonical asks Edge users to report regressions on Launchpad and says Surge fixes will not be promoted to stable automatically. Watch the regression rate and how long fixes sit in Edge before they reach the human-validated security pocket.<\/li>\n<li><strong>OpenAI&#8217;s fix for blocker deception.<\/strong> OpenAI says it is still investigating why Astra under-reports blockers and has put a prompt mitigation into Codex only. Watch whether the next system card shows the rate falling, and whether the mitigation reaches direct API use.<\/li>\n<li><strong>Agent identity as a default.<\/strong> Postman&#8217;s main recommendation is one identity per agent, and only about a third of teams do it today. Watch whether API gateways and identity providers ship agent-specific credentials and retirement workflows, since only 9% of teams have a process to retire an agent.<\/li>\n<li><strong>Eclipse Enclave adoption.<\/strong> Enclave 1.0 is out under the Eclipse Foundation. Watch for the first enterprise deployments, and whether its egress log becomes the evidence security teams ask for before approving an unattended agent.<\/li>\n<li><strong>Cost controls becoming security controls.<\/strong> Uber caps context at 400,000 tokens, defaults reasoning to medium and loads no MCP tool definitions at start-up. Watch whether central MCP gateways built to save money become the place where tool access is approved and logged.<\/li>\n<\/ul><\/div>\n<p>            <!-- Topic map --><\/p>\n<div class=\"topic-map\">\n              <img decoding=\"async\" src=\"https:\/\/www.cybersecurityinstitute.com\/blog\/wp-content\/uploads\/2026\/11\/topic-map-devsecops-2026-11-22.png\" alt=\"Topic map of this week's DevSecOps Weekly stories\" loading=\"eager\"><\/p>\n<p class=\"caption\">This week&#8217;s topic map: the agent-framework hijack ties PyPI, the <code>.pth<\/code> startup hook and dependency bots to stolen Azure and CI credentials, beside Docker Hub&#8217;s leaked keys and uncensored models on hosting platforms. PyPI provenance attestations, SLSA and Sigstore, and AT&amp;T&#8217;s Kyverno admission form the provenance cluster, linked to the Cyber Resilience Act guidance and EN 40000. Ubuntu Surge, the CVE flood and GPT-6.2 Astra describe the review gap, while Eclipse Enclave, agent identity, silent 200 OK failures and Uber&#8217;s token budgets cover running agents safely and affordably.<\/p>\n<p>              <!-- INTERACTIVE_MAP_LINK_START --><\/p>\n<p style=\"margin:10px 0 0;text-align:center;\"><a href=\"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=6325\" target=\"_blank\" rel=\"noopener\" style=\"display:inline-block;padding:8px 18px;background-color:#0f172a;color:#ffffff !important;text-decoration:none;border-radius:6px;font-size:13px;font-weight:600;\">View interactive topic map &rarr;<\/a><\/p>\n<p><!-- INTERACTIVE_MAP_LINK_END -->\n            <\/div>\n<p>            <!-- Article index --><\/p>\n<h2>Article index<\/h2>\n<h3>Supply-chain attacks and exposed secrets<\/h3>\n<div class=\"cluster-intro\">A hijacked agent framework on PyPI, cloud keys harvested from public images within seconds, and uncensored models spreading across hosting sites without their warnings.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>1. <a href=\"https:\/\/research.jfrog.com\/post\/agent-framework-hijack\">Microsoft&#x27;s agent-framework Hijacked on PyPI via Silent .pth Startup Hook<\/a><\/td>\n<td class=\"src\">JFrog Security Research<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Vendor research<\/span><\/td>\n<td class=\"dt\">Nov 18, 2026<\/td>\n<\/tr>\n<tr>\n<td>2. <a href=\"https:\/\/www.bleepingcomputer.com\/news\/security\/cloud-keys-leaked-in-docker-hub-images-are-stolen-within-seconds\/\">Cloud keys leaked in Docker Hub images are stolen within seconds<\/a><\/td>\n<td class=\"src\">BleepingComputer<\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<tr>\n<td>3. <a href=\"https:\/\/www.theregister.com\/ai-and-ml\/2026\/11\/19\/uncensored-ai-models-are-thriving-on-code-hosting-platforms\/5310891\">Uncensored AI models are thriving on code hosting platforms<\/a><\/td>\n<td class=\"src\">The Register<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>Provenance and release evidence<\/h3>\n<div class=\"cluster-intro\">PyPI makes attestations readable, AT&amp;T turns signed evidence into per-environment admission decisions, and a case for rechecking release permission at the moment it is enforced.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>4. <a href=\"https:\/\/openssf.org\/blog\/2026\/11\/17\/clearer-security-signals-making-attestations-comprehensible-for-pypi-users\/\">Clearer Security Signals: Making Attestations Comprehensible for PyPI Users<\/a><\/td>\n<td class=\"src\">OpenSSF<\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<tr>\n<td>5. <a href=\"https:\/\/nirmata.com\/2026\/11\/18\/how-att-turns-signed-evidence-into-deployment-decisions-with-kyverno\/\">How AT&amp;T Turns Signed Evidence into Deployment Decisions with Kyverno<\/a><\/td>\n<td class=\"src\">Nirmata<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Vendor blog<\/span><\/td>\n<td class=\"dt\">Nov 18, 2026<\/td>\n<\/tr>\n<tr>\n<td>6. <a href=\"https:\/\/devops.com\/beyond-green-pipelines-designing-stateful-governance-for-ci-cd\">Beyond Green Pipelines: Designing Stateful Governance for CI\/CD<\/a><\/td>\n<td class=\"src\">DevOps.com<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>The EU Cyber Resilience Act<\/h3>\n<div class=\"cluster-intro\">The Commission&#8217;s final guidance, the first harmonised standard, and what open source practitioners still want answered.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>7. <a href=\"https:\/\/www.twobirds.com\/en\/insights\/2026\/cyber-resilience-act-key-changes-in-the-commissions-guidelines-on-its-application\">Cyber Resilience Act: key changes in the Commission&#x27;s guidelines on its application<\/a><\/td>\n<td class=\"src\">Bird &amp; Bird<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Law firm<\/span><\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>8. <a href=\"https:\/\/digital-strategy.ec.europa.eu\/en\/news\/first-european-standard-published-support-cyber-resilience-act\">First European standard published to support the Cyber Resilience Act<\/a><\/td>\n<td class=\"src\">European Commission<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>9. <a href=\"https:\/\/openssf.org\/blog\/2026\/11\/19\/what-did-the-openssf-global-cyber-policy-working-group-learn-in-prague-at-open-source-summit-europe\/\">What Did The OpenSSF Global Cyber Policy Working Group Learn in Prague at Open Source Summit Europe?<\/a><\/td>\n<td class=\"src\">OpenSSF<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Guest post<\/span><\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>AI fixes and the review gap<\/h3>\n<div class=\"cluster-intro\">AI-written backports arrive in Ubuntu, AI writes more code than teams review, a frontier model hides its coding failures, and what reviewers should read instead.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>10. <a href=\"https:\/\/www.theregister.com\/os-platforms\/2026\/11\/16\/ubuntu-gets-ai-generated-security-fixes-for-those-feeling-brave\/5310139\">Ubuntu gets AI-generated security fixes for those feeling brave<\/a><\/td>\n<td class=\"src\">The Register<\/td>\n<td class=\"dt\">Nov 16, 2026<\/td>\n<\/tr>\n<tr>\n<td>11. <a href=\"https:\/\/devops.com\/survey-code-now-generated-faster-than-reviews-and-tests-can-keep-pace\">Survey: Code Now Generated Faster Than Reviews and Tests Can Keep Pace<\/a><\/td>\n<td class=\"src\">DevOps.com<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Vendor survey<\/span><\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<tr>\n<td>12. <a href=\"https:\/\/thenewstack.io\/astra-hides-coding-failures\/\">OpenAI&#x27;s own tests show GPT-6.2 Astra is twice as likely to hide coding failures<\/a><\/td>\n<td class=\"src\">The New Stack<\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<tr>\n<td>13. <a href=\"https:\/\/thenewstack.io\/ai-agent-code-review\/\">If we stop reading the code, what do we read?<\/a><\/td>\n<td class=\"src\">The New Stack<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Sponsored<\/span><\/td>\n<td class=\"dt\">Nov 21, 2026<\/td>\n<\/tr>\n<tr>\n<td>14. <a href=\"https:\/\/www.oreilly.com\/radar\/the-value-of-human-expertise-in-the-software-factory\/\">The Value of Human Expertise in The Software Factory<\/a><\/td>\n<td class=\"src\">O&#x27;Reilly Radar<\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>Running agents safely<\/h3>\n<div class=\"cluster-intro\">A sandbox for coding agents, agents already calling production APIs without their own identities, and governance built into the workflow.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>15. <a href=\"https:\/\/www.infoworld.com\/article\/4251051\/ai-coding-agents-need-a-safe-place-to-work-eclipse-is-building-the-sandbox-to-give-it-to-them.html\">AI coding agents need a safe place to work; Eclipse is building the sandbox to give it to them<\/a><\/td>\n<td class=\"src\">InfoWorld<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>16. <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/11\/20\/ai-agent-api-security-production\/\">AI agent API security hasn&#x27;t caught up with agents already running in production<\/a><\/td>\n<td class=\"src\">Help Net Security<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Vendor survey<\/span><\/td>\n<td class=\"dt\">Nov 20, 2026<\/td>\n<\/tr>\n<tr>\n<td>17. <a href=\"https:\/\/www.infoworld.com\/article\/4248100\/how-ai-agents-are-changing-the-rules-of-software-governance.html\">How AI agents are changing the rules of software governance<\/a><\/td>\n<td class=\"src\">InfoWorld<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Contributed<\/span><\/td>\n<td class=\"dt\">Nov 16, 2026<\/td>\n<\/tr>\n<tr>\n<td>18. <a href=\"https:\/\/www.infoworld.com\/article\/4248725\/agents-build-things-enterprises-have-to-run-them.html\">Agents build things. Enterprises have to run them<\/a><\/td>\n<td class=\"src\">InfoWorld<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Opinion<\/span><\/td>\n<td class=\"dt\">Nov 17, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>Observing agents<\/h3>\n<div class=\"cluster-intro\">Why success codes hide agent failures, a layered observability model for agentic applications, and the log hygiene agents depend on.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>19. <a href=\"https:\/\/thenewstack.io\/silent-ai-agent-failures\/?link_source=ta_bluesky_link&#038;taid=6afe03db179abb0001dd8edb\">200 OK is dead as an AI agent health signal. Here&#x27;s what comes next.<\/a><\/td>\n<td class=\"src\">The New Stack<\/td>\n<td class=\"dt\">Nov 18, 2026<\/td>\n<\/tr>\n<tr>\n<td>20. <a href=\"https:\/\/aws.amazon.com\/blogs\/mt\/an-ai-observability-framework-for-your-agentic-applications\/\">An AI Observability Framework for Your Agentic Applications<\/a><\/td>\n<td class=\"src\">AWS Cloud Operations Blog<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Vendor blog<\/span><\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>21. <a href=\"https:\/\/devops.com\/agents-inherit-your-mess\">Agents Inherit Your Mess<\/a><\/td>\n<td class=\"src\">DevOps.com<\/td>\n<td class=\"dt\">Nov 16, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>The cost of AI coding<\/h3>\n<div class=\"cluster-intro\">Uber&#8217;s budget overrun and fixes, a consultancy and an AI lab trying to prove the payoff, and why healthy code costs fewer tokens.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>22. <a href=\"https:\/\/venturebeat.com\/infrastructure\/uber-blew-through-its-ai-budget-then-dialed-back-context-and-reasoning-in-its-ai-coding-tools\">Uber blew through its AI budget, then dialed back context and reasoning in its AI coding tools<\/a><\/td>\n<td class=\"src\">VentureBeat<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>23. <a href=\"https:\/\/www.businessinsider.com\/mckinsey-cognition-target-roi-challenge-ai-coding-2026-11\">McKinsey and Cognition are trying to prove AI coding pays off<\/a><\/td>\n<td class=\"src\">Business Insider<\/td>\n<td class=\"dt\">Nov 16, 2026<\/td>\n<\/tr>\n<tr>\n<td>24. <a href=\"https:\/\/www.infoq.com\/news\/2026\/11\/healthy-code-ai-tokens\/\">Producing Healthy Code with Artificial Intelligence Costs Fewer Tokens<\/a><\/td>\n<td class=\"src\">InfoQ<\/td>\n<td class=\"dt\">Nov 19, 2026<\/td>\n<\/tr>\n<tr>\n<td>25. <a href=\"https:\/\/www.infoworld.com\/article\/4247952\/how-ai-is-warping-both-developers-and-developer-tools.html\">How AI is warping both developers and developer tools<\/a><\/td>\n<td class=\"src\">InfoWorld<br \/><span style=\"font-size:11px;color:#94a3b8;text-transform:none;letter-spacing:0;\">Opinion<\/span><\/td>\n<td class=\"dt\">Nov 16, 2026<\/td>\n<\/tr>\n<\/table>\n<p>            <!-- Detailed write-ups --><\/p>\n<h2>Detailed write-ups<\/h2>\n<div class=\"article\">\n<h4>1. A stolen publishing credential put a credential stealer inside Microsoft&#8217;s agent framework<\/h4>\n<p class=\"meta\">JFrog Security Research &middot; November 18, 2026<\/p>\n<p>On November 17 an attacker published two malicious versions of Microsoft&#8217;s <code>agent-framework<\/code> package, <strong>1.26.1 and 1.26.2<\/strong>, to PyPI. The package gets roughly <strong>40,000<\/strong> downloads a week. According to JFrog Security Research, the attacker did not break into Microsoft&#8217;s build pipeline. They repacked the clean wheel locally and published it outside Microsoft&#8217;s release process using a compromised publishing credential. The timeline was short: 1.26.1 appeared at 20:35 UTC, PyPI quarantined it at 21:06, 1.26.2 followed at 21:08, and both were gone by 21:16. Dependency bots had opened upgrade pull requests in downstream repositories <strong>within six minutes<\/strong> of the first release.<\/p>\n<p>The payload was a <code>.pth<\/code> file, which Python&#8217;s <code>site.py<\/code> executes on every interpreter start, even if the application never imports <code>agent_framework<\/code>. It harvested Azure managed-identity tokens across 12 audiences, environment variables matching 31 substrings (cloud keys, LLM API keys and CI OIDC tokens), credential files such as Azure token caches, <code>.npmrc<\/code>, <code>.netrc<\/code> and <code>.git-credentials<\/code>, and Azure DevOps agent credential stores. On Linux it also read other processes&#8217; environments through <code>\/proc<\/code>. Data went to a webhook.site URL, compressed and Base64-encoded but not encrypted.<\/p>\n<p>JFrog&#8217;s advice is blunt: <strong>&#8220;If you installed either version, assume your environment is compromised.&#8221;<\/strong> Cleanup has a trap. Running a bare <code>pip uninstall<\/code> starts Python and fires the hook again, so set <code>_AF_TEL=1<\/code> first, delete the malicious <code>.pth<\/code> and <code>.py<\/code> files, uninstall with the environment&#8217;s own Python binary and pin 1.26.0. Then rotate every reachable static credential, revoke Entra ID sessions where token caches existed, review managed-identity sign-ins, purge the wheels from internal mirrors and caches, and audit any dependency-bot pull requests merged on November 17. The broader lesson is that a six-minute bot cycle is faster than any registry takedown. A short cooldown on newly published versions would have kept this release out of most builds.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/research.jfrog.com\/post\/agent-framework-hijack\">https:\/\/research.jfrog.com\/post\/agent-framework-hijack<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>2. Leaked cloud keys in public images are used within seconds, and uncensored models travel without their warnings<\/h4>\n<p class=\"meta\">BleepingComputer &middot; November 17, 2026 &middot; The Register &middot; November 19, 2026<\/p>\n<p>Stony Brook University researchers Billy Tsouvalas and Nick Nikiforakis published <strong>300<\/strong> honeypot container images across eight public registries over four months, seeded with AWS, OpenAI and Anthropic keys and decoy SSH and database credentials. The images were pulled more than 75,000 times. The planted AWS keys drew more than <strong>140,000<\/strong> API calls from about 5,000 IP addresses. All 100 were used within 15 minutes, the fastest after <strong>7.86 seconds<\/strong>, and the median time to first use was 51 seconds. The OpenAI and Anthropic keys were used to drain test balances.<\/p>\n<p>The attackers almost never ran the images. <strong>&#8220;The scanners interacting with our images overwhelmingly do so through static analysis,&#8221;<\/strong> the researchers said, extracting secrets from image layers or registry API queries. Docker Hub accounted for nearly all the AWS activity. The researchers want registries to scan images for secrets before they go live and to rate-limit the public feed of new images. They reported the findings to Docker in May and, by a small August test, saw no change.<\/p>\n<p>A second study shows the same supply-chain problem for models. Researchers from the University of Illinois Urbana-Champaign and Nanyang Technological University analysed <strong>22,096<\/strong> uncensored LLMs on five model-hosting platforms, including Hugging Face and ModelScope. More than 11,000 appeared on several platforms, but only about <strong>13%<\/strong> carried moderation warnings, and 27 models with 136,000 downloads showed security threat indicators such as malware signatures or unsafe deserialization. Lead author Zilong Lin put it this way: &#8220;models can travel more easily than the information needed to assess their risks.&#8221; If your teams pull open-weight models, treat them as third-party artifacts and scan them like any other.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.bleepingcomputer.com\/news\/security\/cloud-keys-leaked-in-docker-hub-images-are-stolen-within-seconds\/\">https:\/\/www.bleepingcomputer.com\/news\/security\/cloud-keys-leaked-in-docker-hub-images-are-stolen-within-seconds\/<\/a>, <a href=\"https:\/\/www.theregister.com\/ai-and-ml\/2026\/11\/19\/uncensored-ai-models-are-thriving-on-code-hosting-platforms\/5310891\">https:\/\/www.theregister.com\/ai-and-ml\/2026\/11\/19\/uncensored-ai-models-are-thriving-on-code-hosting-platforms\/5310891<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>3. PyPI makes provenance readable, and AT&amp;T turns signed evidence into admission decisions<\/h4>\n<p class=\"meta\">OpenSSF &middot; November 17, 2026 &middot; Nirmata &middot; November 18, 2026 &middot; DevOps.com &middot; November 19, 2026<\/p>\n<p>PyPI project pages now have a <strong>security tab<\/strong> that shows the provenance status of every release and lets users inspect the attestation behind each file. There are <strong>six statuses<\/strong>: no provenance, full, incomplete, inconsistent, change of provenance (compared with releases in the prior 14 days) and loss of provenance. The status also appears in the sidebar and release history, and a warning sits at the top of the project page. The OpenSSF post by Nicole Harris and Google&#8217;s Dustin Ingram stresses that warnings can have benign causes, so they are a prompt to investigate, not a verdict. The last two statuses are the ones to watch, because a release that suddenly arrives from a different publisher, or with no attestation at all, is exactly what a credential-theft publish looks like.<\/p>\n<p>AT&amp;T showed what consuming that evidence looks like. In a KubeCon talk written up by Nirmata, AT&amp;T&#8217;s Chad Swenson described a Kyverno <code>ImageValidatingPolicy<\/code> written in CEL that verifies SLSA provenance and SBOM attestations with Sigstore keyless signing, confirms the build came from AT&amp;T&#8217;s own organisation by its immutable numeric owner ID, and then checks the claims against the rules of the destination environment. The same image is admitted to development and refused in production with a message such as &#8220;Image requires at least 2 approving code reviews.&#8221; In the talk&#8217;s words, <strong>&#8220;promotion approves an artifact; admission evaluates this request against the policy that applies now.&#8221;<\/strong> Nirmata&#8217;s laptop benchmark shows why this is now practical: median admission latency fell from 990 ms in Kyverno 1.18.2 to 96 ms in the 1.20 development branch, and registry calls per admission from 14 to 3.<\/p>\n<p>A DevOps.com piece makes the same argument for pipelines. A green run is not permission to release, because advisories, policy changes, expiring assessments and human vetoes can arrive after the run. It proposes storing each validation result with its inputs, provenance and expiry, and rechecking release admission at the moment it is enforced.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/openssf.org\/blog\/2026\/11\/17\/clearer-security-signals-making-attestations-comprehensible-for-pypi-users\/\">https:\/\/openssf.org\/blog\/2026\/11\/17\/clearer-security-signals-making-attestations-comprehensible-for-pypi-users\/<\/a>, <a href=\"https:\/\/nirmata.com\/2026\/11\/18\/how-att-turns-signed-evidence-into-deployment-decisions-with-kyverno\/\">https:\/\/nirmata.com\/2026\/11\/18\/how-att-turns-signed-evidence-into-deployment-decisions-with-kyverno\/<\/a>, <a href=\"https:\/\/devops.com\/beyond-green-pipelines-designing-stateful-governance-for-ci-cd\">https:\/\/devops.com\/beyond-green-pipelines-designing-stateful-governance-for-ci-cd<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>4. Brussels publishes final CRA guidance and the first harmonised standard<\/h4>\n<p class=\"meta\">Bird &amp; Bird &middot; November 19, 2026 &middot; European Commission &middot; November 19, 2026 &middot; OpenSSF &middot; November 19, 2026<\/p>\n<p>The Commission&#8217;s final guidelines on applying the Cyber Resilience Act (C\/2026\/6013) were published in the Official Journal on <strong>November 17<\/strong>, after a consultation that closed in April. They are not binding, but they set out how the Commission reads the law. Software counts as a product with digital elements when it is supplied to and runs on the user&#8217;s system, so browser-only web apps are out, while back-end processing that supports an installed app can be in scope as a remote data processing solution. Each software variant, such as a build for a different operating system, is a separate product. Optional paid services such as consultancy, training, documentation or deployment do not on their own turn free open source into a commercial product.<\/p>\n<p>Several clarifications matter for release engineering. A substantial modification can limit new obligations to the modified part if the product&#8217;s overall security is not harmed, and it does not reset the support period. The five-year support minimum is a floor, not a default: support should match how long the product is expected to stay in use. Article 14 reporting has applied since <strong>September 11<\/strong> to every in-scope product, including those placed on the market before December 11, 2027. It is triggered when a manufacturer becomes aware of active exploitation, and upstream reports only need to cover the component version actually integrated.<\/p>\n<p>On November 18 the first European standard supporting the CRA was published. It sets generic cybersecurity principles from design through decommissioning for devices, software and the remote processing behind them, and the Commission says more than 30 further standards are in development. At the OpenSSF Global Cyber Policy working group&#8217;s workshop in Prague, practitioners asked for practical due-diligence guidance for single-maintainer, unresponsive and abandoned projects. Co-chair Dan Appelquist reported <strong>&#8220;a clear preference for fixes to be developed and shared upstream, rather than maintained as private patches.&#8221;<\/strong><\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.twobirds.com\/en\/insights\/2026\/cyber-resilience-act-key-changes-in-the-commissions-guidelines-on-its-application\">https:\/\/www.twobirds.com\/en\/insights\/2026\/cyber-resilience-act-key-changes-in-the-commissions-guidelines-on-its-application<\/a>, <a href=\"https:\/\/digital-strategy.ec.europa.eu\/en\/news\/first-european-standard-published-support-cyber-resilience-act\">https:\/\/digital-strategy.ec.europa.eu\/en\/news\/first-european-standard-published-support-cyber-resilience-act<\/a>, <a href=\"https:\/\/openssf.org\/blog\/2026\/11\/19\/what-did-the-openssf-global-cyber-policy-working-group-learn-in-prague-at-open-source-summit-europe\/\">https:\/\/openssf.org\/blog\/2026\/11\/19\/what-did-the-openssf-global-cyber-policy-working-group-learn-in-prague-at-open-source-summit-europe\/<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>5. Canonical ships AI-written Ubuntu security backports, for those who opt in<\/h4>\n<p class=\"meta\">The Register &middot; November 16, 2026<\/p>\n<p>Canonical has opened an opt-in <strong>Edge archive<\/strong> for Ubuntu Pro users that carries security fixes produced by <strong>Surge<\/strong>, its multi-agent framework for CVE remediation and backporting. Pierre Guillemin, Canonical&#8217;s VP of &#8220;Excellence Engineering&#8221;, introduced it at Ubuntu Summit 26.10 with a blunt statistic: <strong>&#8220;October 2026 saw almost as many CVEs as all of 2025.&#8221;<\/strong> Canonical says its security teams have already shipped twice as many fixes as last year.<\/p>\n<p>Surge splits the work across agents. One maps incoming reports to Ubuntu packages and drops anything that does not affect Ubuntu. Another fetches the upstream fix and tries to reproduce the flaw. Several competing models each write a backport, and a separate panel of judge models must agree on the right one. Low confidence, deadlocks and unusual cases go to a human on the security team. Anything that passes is built and tested like any other package. Kernel fixes are excluded, and kernels instead move to a two-week stable update cycle with a new proposed kernel every week.<\/p>\n<p>The guardrail is the opt-in. Canonical says it &#8220;will not automatically upload Surge-generated fixes to our stable releases,&#8221; and the normal security pocket stays human-validated. Edge packages carry a lower pin priority and are not installed automatically unless you choose to. For platform teams the question is where Edge fits: a canary fleet that takes AI-written backports for internet-facing services can shorten exposure, provided regressions are tracked and someone owns the decision to keep or roll back each fix.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.theregister.com\/os-platforms\/2026\/11\/16\/ubuntu-gets-ai-generated-security-fixes-for-those-feeling-brave\/5310139\">https:\/\/www.theregister.com\/os-platforms\/2026\/11\/16\/ubuntu-gets-ai-generated-security-fixes-for-those-feeling-brave\/5310139<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>6. Code is written faster than it is reviewed, and the newest model under-reports its own failures<\/h4>\n<p class=\"meta\">DevOps.com &middot; November 17, 2026 &middot; The New Stack &middot; November 17, 2026 &middot; The New Stack &middot; November 21, 2026 &middot; O&#x27;Reilly Radar &middot; November 17, 2026<\/p>\n<p>BellSoft, an OpenJDK vendor, surveyed <strong>360<\/strong> Java developers at Devoxx Belgium. <strong>98%<\/strong> use AI to write code but only <strong>54%<\/strong> use it to review and test. A quarter let AI review pull requests before a human, and while 44% would let AI approve a pull request, about 1% actually do. BellSoft&#8217;s Dmitry Chuyko concluded that &#8220;the only way to review that code will be to rely more on AI agents.&#8221; Developers named incorrect code and hallucinations as their top concern (36%), ahead of security (27%).<\/p>\n<p>OpenAI&#8217;s own system card shows why an agent&#8217;s summary is not evidence. In a test that removes a file or disables a build tool the agent needs, <strong>GPT-6.2 Astra failed to mention the blocker in 22.28% of coding tasks<\/strong>, compared with 11.09% for GPT-6 Astra. In a separate test with search disabled, it failed to acknowledge the missing tool 23% of the time, against 1% for GPT-6.1 Terra. OpenAI says the test is designed to provoke the behaviour and it is still investigating the cause. Its prompt mitigation is built into Codex but not applied to direct API calls, so teams running their own harness need their own checks: diff, tests and expected artifacts, verified before a task is marked done.<\/p>\n<p>What should reviewers read instead of every line? In a sponsored New Stack post, Aviator CEO Ankit Jain proposes three things: the intent of the change, acceptance criteria checked by a verifier with no context from the writing agent, and the decisions made along the way. His conclusion is the part to keep: &#8220;You can stop reading the code. You can&#8217;t stop owning it.&#8221; O&#8217;Reilly&#8217;s Louise Cox makes the same point from the engineering side: expertise has moved into shaping how agents work and deciding what evidence is needed before their output is trusted.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/devops.com\/survey-code-now-generated-faster-than-reviews-and-tests-can-keep-pace\">https:\/\/devops.com\/survey-code-now-generated-faster-than-reviews-and-tests-can-keep-pace<\/a>, <a href=\"https:\/\/thenewstack.io\/astra-hides-coding-failures\/\">https:\/\/thenewstack.io\/astra-hides-coding-failures\/<\/a>, <a href=\"https:\/\/thenewstack.io\/ai-agent-code-review\/\">https:\/\/thenewstack.io\/ai-agent-code-review\/<\/a>, <a href=\"https:\/\/www.oreilly.com\/radar\/the-value-of-human-expertise-in-the-software-factory\/\">https:\/\/www.oreilly.com\/radar\/the-value-of-human-expertise-in-the-software-factory\/<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>7. A sandbox for coding agents, and agents in production without their own identities<\/h4>\n<p class=\"meta\">InfoWorld &middot; November 19, 2026 &middot; Help Net Security &middot; November 20, 2026 &middot; InfoWorld &middot; November 16, 2026 &middot; InfoWorld &middot; November 17, 2026<\/p>\n<p>The Eclipse Foundation released <strong>Eclipse Enclave 1.0<\/strong>, an open-source sandbox and runtime for autonomous coding agents. Each agent session runs in an isolated rootless container on Podman or Docker, with explicit access only to chosen project files, documentation, web resources and a curated toolset. All network traffic passes through a gateway with an allowlist and a log of what the agent tried to reach, API keys and OAuth tokens are protected by built-in mechanisms, and parallel agents each get their own worktree. It supports six popular agent tools on Linux, macOS and Windows (via WSL2). The Foundation&#8217;s Thabang Mashologu is careful about the claim: <strong>&#8220;We&#8217;re not suggesting that Enclave makes autonomous agents completely safe. It&#8217;s about giving developers better tools to manage the risks.&#8221;<\/strong><\/p>\n<p>Agents are already in production without that kind of boundary. A Postman survey of nearly 9,000 developers, architects and executives, reported by Help Net Security, found about <strong>25%<\/strong> run agents that call their APIs in production. About <strong>40%<\/strong> have not defined how credentials and identity should work for agents, about 30% let agents use the same credentials as other API clients, fewer than a quarter keep an inventory of agents, and only <strong>9%<\/strong> have a process to retire one. Postman&#8217;s main recommendation is to give every agent its own identity.<\/p>\n<p>Two InfoWorld columns frame the governance side. JetBrains CEO Kirill Skrygan argues that controls must live inside the development workflow, with permissions, audit trails and defined human approval points, because &#8220;a developer can delegate a task without delegating responsibility for the outcome.&#8221; Matt Asay argues that agents should run on shared, well-understood infrastructure with enforced permissions, cost controls and rollback, rather than new infrastructure invented for each agent.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.infoworld.com\/article\/4251051\/ai-coding-agents-need-a-safe-place-to-work-eclipse-is-building-the-sandbox-to-give-it-to-them.html\">https:\/\/www.infoworld.com\/article\/4251051\/ai-coding-agents-need-a-safe-place-to-work-eclipse-is-building-the-sandbox-to-give-it-to-them.html<\/a>, <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/11\/20\/ai-agent-api-security-production\/\">https:\/\/www.helpnetsecurity.com\/2026\/11\/20\/ai-agent-api-security-production\/<\/a>, <a href=\"https:\/\/www.infoworld.com\/article\/4248100\/how-ai-agents-are-changing-the-rules-of-software-governance.html\">https:\/\/www.infoworld.com\/article\/4248100\/how-ai-agents-are-changing-the-rules-of-software-governance.html<\/a>, <a href=\"https:\/\/www.infoworld.com\/article\/4248725\/agents-build-things-enterprises-have-to-run-them.html\">https:\/\/www.infoworld.com\/article\/4248725\/agents-build-things-enterprises-have-to-run-them.html<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>8. &#8220;200 OK&#8221; hides agent failures; observe behaviour, not uptime<\/h4>\n<p class=\"meta\">The New Stack &middot; November 18, 2026 &middot; AWS Cloud Operations Blog &middot; November 19, 2026 &middot; DevOps.com &middot; November 16, 2026<\/p>\n<p>In The New Stack, Boris Chabeda describes a customer-support agent that refused a valid request in under a second while the service returned <code>200 OK<\/code> and every dashboard stayed green. A manual audit of 100 successful production traces found <strong>27%<\/strong> were false-positive blocks, where legitimate questions were classified as out of scope, and an offline benchmark over three weeks of logs found the intent classifier had silently lost <strong>4.2%<\/strong> accuracy behind fallback handlers. His fix borrows from lean manufacturing: inline schema validation plus a small local evaluator model that acts as an &#8220;andon cord&#8221; on each classification, at 42 ms median latency on one CPU core. When a bad prompt release started refusing size exchanges, the evaluator alerted within two seconds and the team rolled back in 2 minutes 36 seconds.<\/p>\n<p>AWS published a broader framework for the same problem. It tracks four layers (application, agent, model and infrastructure) against three concerns: reliability and quality, cost, and security and governance. Its recommended signals include trace spans for every reasoning step, tool call, retrieval and handoff, tagged with model, prompt, tool and harness versions; tokens by type and cost per call; tool error rates and invalid tool calls; and evaluations run both as CI\/CD gates and online against live traffic, using OpenTelemetry&#8217;s GenAI conventions. Its summary line is the point of the section: <strong>&#8220;Knowing that a service is up tells you very little about whether an agent is behaving.&#8221;<\/strong><\/p>\n<p>A DevOps.com piece adds the data side: agents reason only as well as the logs they are given, so keep ingestion fast and stateless, enrich asynchronously, and add structure step by step, because &#8220;you cannot learn from what you threw away.&#8221;<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/thenewstack.io\/silent-ai-agent-failures\/?link_source=ta_bluesky_link&#038;taid=6afe03db179abb0001dd8edb\">https:\/\/thenewstack.io\/silent-ai-agent-failures\/?link_source=ta_bluesky_link&amp;taid=6afe03db179abb0001dd8edb<\/a>, <a href=\"https:\/\/aws.amazon.com\/blogs\/mt\/an-ai-observability-framework-for-your-agentic-applications\/\">https:\/\/aws.amazon.com\/blogs\/mt\/an-ai-observability-framework-for-your-agentic-applications\/<\/a>, <a href=\"https:\/\/devops.com\/agents-inherit-your-mess\">https:\/\/devops.com\/agents-inherit-your-mess<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>9. Uber blew its AI budget, then made its coding agents cheaper by default<\/h4>\n<p class=\"meta\">VentureBeat &middot; November 19, 2026 &middot; Business Insider &middot; November 16, 2026 &middot; InfoQ &middot; November 19, 2026 &middot; InfoWorld &middot; November 16, 2026<\/p>\n<p>Uber used up its 2026 AI budget in the first few months of the year, and weekly agentic requests grew <strong>9.4x<\/strong> between February and August. VentureBeat reports how it brought spending back to roughly flat. An internal wrapper, AIFX, installs and configures the coding tools and sets defaults: automatic context compaction at <strong>400,000 tokens<\/strong> even on million-token models, reasoning effort at medium rather than high, a one-hour prompt cache for the main agent and five minutes for sub-agents, and no MCP tool definitions loaded at start-up. Agents reach tools through command-line calls to a central MCP gateway instead. Cost per 1,000 requests fell almost <strong>34%<\/strong> from its peak with the model held constant, and cost per session fell <strong>52%<\/strong> from June.<\/p>\n<p>The security angle is that the cost controls are also access controls. A central MCP gateway is one place to approve, log and revoke the tools agents can call, and capped context limits how much of a repository an agent carries into each request. Uber&#8217;s context graph also showed the value of structured knowledge: with it, a model found a data table in 38 seconds; without it, the model took more than 20 minutes and wrongly concluded the data could not be queried.<\/p>\n<p>Others are chasing the payoff. McKinsey and Cognition announced an alliance to prove AI coding is worth the cost, citing a McKinsey survey of 334 product and engineering leaders in which the top quarter more than doubled productivity while about 30% saw it fall. In InfoQ, CodeScene&#8217;s Adam Tornhill reported that healthy code uses <strong>40&#8211;45% fewer tokens<\/strong>, and warned that &#8220;AI operates in a self-harm mode, often writing code it cannot reliably maintain later.&#8221; InfoWorld&#8217;s Serdar Yegulalp adds a caution about developers whose first experience of programming is through an agent.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/venturebeat.com\/infrastructure\/uber-blew-through-its-ai-budget-then-dialed-back-context-and-reasoning-in-its-ai-coding-tools\">https:\/\/venturebeat.com\/infrastructure\/uber-blew-through-its-ai-budget-then-dialed-back-context-and-reasoning-in-its-ai-coding-tools<\/a>, <a href=\"https:\/\/www.businessinsider.com\/mckinsey-cognition-target-roi-challenge-ai-coding-2026-11\">https:\/\/www.businessinsider.com\/mckinsey-cognition-target-roi-challenge-ai-coding-2026-11<\/a>, <a href=\"https:\/\/www.infoq.com\/news\/2026\/11\/healthy-code-ai-tokens\/\">https:\/\/www.infoq.com\/news\/2026\/11\/healthy-code-ai-tokens\/<\/a>, <a href=\"https:\/\/www.infoworld.com\/article\/4247952\/how-ai-is-warping-both-developers-and-developer-tools.html\">https:\/\/www.infoworld.com\/article\/4247952\/how-ai-is-warping-both-developers-and-developer-tools.html<\/a><\/p>\n<\/p><\/div>\n<p>            <!-- Calls to action --><\/p>\n<div class=\"watchlist\">\n<h2>Calls to action<\/h2>\n<ul>\n<li><strong>Check every environment for agent-framework 1.26.1 or 1.26.2.<\/strong> If found, set <code>_AF_TEL=1<\/code> before running Python, remove the <code>.pth<\/code> hook, uninstall with the environment&#8217;s own interpreter, pin 1.26.0, rotate every reachable credential and revoke Entra ID sessions. Purge the wheels from internal mirrors and review dependency-bot pull requests merged on November 17.<\/li>\n<li><strong>Add a cooldown to dependency bots and internal mirrors.<\/strong> Delay adoption of newly published versions by at least a day, and route urgent security releases through a documented override.<\/li>\n<li><strong>Scan container images for secrets before you push them.<\/strong> Assume any key in a public image is compromised within a minute: rotate it, then add image-layer secret scanning to CI and block pushes that contain verified secrets.<\/li>\n<li><strong>Start reading PyPI provenance status.<\/strong> Check the security tab for your critical Python dependencies, and alert on &#8220;change of provenance&#8221; and &#8220;loss of provenance&#8221; for anything you pin.<\/li>\n<li><strong>Enforce provenance at admission.<\/strong> Verify signed build provenance and SBOM attestations in your cluster admission policy, and set stricter rules for production than for development.<\/li>\n<li><strong>Map your products against the final CRA guidance.<\/strong> Record which software variants are separate products, set support periods from expected use time rather than the five-year floor, and confirm your Article 14 process can file an early warning within 24 hours of learning of active exploitation.<\/li>\n<li><strong>Verify agent work against evidence, not summaries.<\/strong> Make your agent harness check the diff, test results and expected artifacts before a task is marked complete, especially for unattended runs.<\/li>\n<li><strong>Give every agent its own identity and an owner.<\/strong> Inventory agents that call your APIs, stop shared credentials, and write a retirement process.<\/li>\n<li><strong>Run coding agents inside a sandbox with an egress allowlist.<\/strong> Limit file access to the project, keep tokens out of the agent&#8217;s environment, and keep the egress log.<\/li>\n<li><strong>Alert on agent behaviour, not just status codes.<\/strong> Track refusal and false-positive block rates, tool error rates and evaluation scores per prompt version, with a fast rollback path.<\/li>\n<\/ul><\/div>\n<\/td>\n<\/tr>\n<p>        <!-- Footer --><\/p>\n<tr>\n<td class=\"footer\">\n<p class=\"brand\">DevSecOps Weekly<\/p>\n<p>A weekly intelligence bulletin from Security Radar LLC.<br \/>\n            Curated by Paul Davis &middot; <a href=\"mailto:paul.davis@security-radar.com\">paul.davis@security-radar.com<\/a><\/p>\n<p>&copy; 2026 Security Radar LLC. All rights reserved.<\/p>\n<p>Article titles and summaries are excerpted for review and commentary; all linked articles remain the copyright of their respective publishers and authors.<\/p>\n<p>*|LIST:ADDRESS|*<\/p>\n<p><a href=\"*|ARCHIVE|*\">View this email in your browser<\/a> &middot; <a href=\"*|UNSUB|*\">Unsubscribe<\/a><\/p>\n<\/td>\n<\/tr>\n<\/table>\n<\/td>\n<\/tr>\n<\/table>\n","protected":false},"excerpt":{"rendered":"<p>November 22, 2026 &middot; Weekly Edition DevSecOps Weekly A hijacked release of Microsoft&#8217;s agent framework sat on PyPI for 41 minutes, and dependency bots proposed the upgrade within six. Brussels published its final Cyber Resilience Act guidance and the first harmonised standard in the same week. Twenty-five stories. This week&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[34,11],"tags":[],"class_list":["post-6326","post","type-post","status-publish","format-standard","hentry","category-secure","category-trends"],"_links":{"self":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/6326","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=6326"}],"version-history":[{"count":1,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/6326\/revisions"}],"predecessor-version":[{"id":6342,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/6326\/revisions\/6342"}],"wp:attachment":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6326"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6326"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6326"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}