Five things from the past working week that a UK board should know about, in the order I would raise them with a chair over coffee on Monday morning. Microsoft shipped more than a thousand fixes for the second month running, a stolen-token campaign worked its way through thousands of build pipelines, Maersk was breached through a company it bought four years ago, the FCA told trading venues what their boards were not doing, and the NCSC started vouching for the people you pay to attack you. I have left out the noise. What survives has a decision attached.
1. Microsoft fixed more than a thousand flaws, and the volume is now the story
November's Patch Tuesday on 10 November listed 1,094 CVEs, 726 of them in Windows and 174 in Office, the second consecutive month above a thousand. One is already being exploited: CVE-2026-124194, an Outlook flaw rated only CVSS 6.5, which CISA added to its Known Exploited Vulnerabilities catalogue shortly after release. The Register explains that a crafted message file makes Outlook connect to an attacker's server without the usual warning, handing over the user's NTLMv2 hash, which can then be used to sign in as that person. Three further bugs carry a 9.8 rating, including CVE-2026-116928 in Windows Search, which can be triggered through Exchange by an attachment that is indexed automatically and never opened, and CVE-2026-119958 in Windows LDAP, which puts domain controllers in reach. Trend Micro's Dustin Childs puts the volume down to automated bug-hunting and says AI makes reversing a patch into an exploit close to instant. The low score on the exploited bug is the trap: severity ratings are not a substitute for knowing what is being used against people now.
For boards. Ask whether your patching targets were set when a month meant a hundred fixes, how long it currently takes to get an exploited flaw onto every Outlook client and domain controller, and who decides what waits.
2. A stolen-token campaign ran through roughly 1,800 build pipelines in three days
Between 10 and 12 November an attacker used stolen GitHub credentials to plant a secret-stealing script in code repositories those accounts could write to. Semgrep's analysis counts 4,837 accounts that opened probe pull requests into 61,764 repositories, and at least 1,800 repositories where the payload ran, a figure it expects to rise. The Hacker News reports that the script harvested CI secrets and cloud credentials, sent them out over HTTP or Telegram, and then reverted its own changes within minutes, often deleting the run logs. Accounts caught in October's GhostAction wave were far more likely to be reused here, which tells you something about how many organisations rotated nothing after the last one. There is no UK victim named yet. That is not the same as there being none.
For boards. Ask whether your engineering teams, and your software suppliers, can show that every credential exposed in October's incident was rotated, and whether a build pipeline that quietly changed and changed back would be noticed by anyone.
3. Maersk was breached through a company it bought in 2022
On 11 November Maersk confirmed an intruder had extracted data from a single legacy system inherited through an acquisition, saying its findings "strongly suggest no sensitive data has been disclosed" and that systems connected by API and EDI were unaffected. BleepingComputer identifies the system as the OCF platform from LF Logistics, acquired from Li & Fung in 2022. A new extortion group, WhiteArmor, claims around 100GB including driver records, plaintext passwords and 33 super-administrator accounts, and wants $5 million in thirty days. Those claims are unverified. The structural point is not: four years after completion, a system nobody would choose to build today was still reachable and still holding data worth stealing.
For boards. Ask for a list of every system you inherited through acquisition that is still running, who owns each one, and the date it is due to be switched off. If no one can produce that list within a week, that is the answer.
4. The FCA found trading venue boards could not show they had challenged resilience plans
On 11 November the FCA published a multi-firm review of operational resilience at UK trading venues, covering MTFs, OTFs and recognised investment exchanges. Some boards had limited evidence of reviewing, challenging or approving resilience work and remediation plans. Where a venue sits in a group, the UK entity remains responsible for its UK activities even when group teams run the services, and some had not defined that accountability. Third-party dependencies and concentration risk were poorly understood, exit plans were often high-level, and the FCA notes that contract clauses alone do not guarantee recovery. It is not introducing new requirements, but Norton Rose Fulbright's summary highlights that the FCA's operational incident reporting rules take effect on 18 March 2027. The findings read across to most regulated firms, not just venues.
For boards. Ask whether your minutes show the board challenging resilience plans rather than noting them, and whether any of your important business services depends on a group service whose UK accountability is assumed rather than written down.
5. The NCSC will now vouch for the people you pay to attack you
On 11 November the NCSC launched its Cyber Adversary Simulation scheme, which assures commercial providers who simulate realistic attacks against large organisations, critical national infrastructure and government. Providers are assessed against an NCSC standard and are expected to build attack plans specific to the customer rather than run a fixed menu of techniques. The scheme sits alongside the Cyber Assessment Framework, and regulators and lead government departments can use it to gather evidence across their sectors. It launches as a minimum viable product, with assured firms listed on the CyAS scheme pages. "You cannot improve what you do not know," said the NCSC's CTO, Ollie Whitehouse. The subtext for anyone in scope of the Cyber Security and Resilience Bill is that this is how a regulator may one day ask you to prove it.
For boards. Ask when your defences were last tested by someone behaving like an attacker rather than an auditor, what they got to, and whether the next test should be bought from the assured list.
The thread that ties this together
Every story this week was about something inherited. Tokens stolen in October were spent in November. Maersk's exposure arrived with an acquisition. Trading venues leaned on group services whose accountability nobody had written down. Microsoft's patch volume is the backlog of years of code now being searched by machines. The Cyber Security and Resilience Bill returns to the Commons on 24 November to consider the Lords amendments, and when it lands it will ask regulated firms to evidence resilience, not assert it.
What has your organisation inherited, whether through a deal, a supplier or last month's incident, that nobody has yet checked is still safe to keep?