TfL pulls Oyster photocard site offline after account breach hits 265,000
Account details of around 265,000 photocard holders, including children as young as five, may have been compromised. Passwords and payment details were reportedly untouched.
CRUNCHY TECH
EVERY PUBLISHED BRIEF
Browse Crunchy Tech’s latest security briefs, grouped by date.
PUBLISHED
Account details of around 265,000 photocard holders, including children as young as five, may have been compromised. Passwords and payment details were reportedly untouched.
Two CVSS 9.8 bugs are under attack, one via Confluence Data Center, and Apache will not fix the end-of-life Axis 1.x.
The DOJ alleges Igor D'iu was recruited via Telegram to surveil employees; Bloomberg reports the assignment stemmed from Ukraine's wartime use of the firm's software.
PUBLISHED
F5 has shipped emergency fixes for 10 NGINX bugs, including two unauthenticated remote code execution flaws rated 9.2 that need HTTP/2 enabled.
Mandiant researchers describe how Cold Relic turned data from one breached US organisation into convincing impersonation attacks on a second.
A bipartisan Senate letter asks Jay Clayton to secure US systems after nationwide outages, amid claims that AI finds software flaws faster than organisations can patch them.
Public preview locks down Server Core from first boot, and Microsoft warns some things that worked on Server 2025 may break.
A 9.4-rated remote code execution bug in Puppet Enterprise and Puppet Core is being exploited, and CISA has given US federal agencies two weeks to patch.
A memory safety bug in Base64 decoding, tracked as CVE-2026-114938, allows unauthenticated code execution on popular enterprise software, though researchers say real-world exploitation takes per-target work.
PUBLISHED
Malicious JavaScript on 178 websites skimmed payment cards, swapped crypto wallet addresses and hit iPhones and iPads with exploit kits, before Optimizely blocked the vulnerable path.
Reports to NCMEC climbed from about 4,700 AI-generated abuse images in 2023 to over 400,000 in 2025, according to Edinburgh's Childlight Institute.
NATO agencies and several European governments can now use OpenAI's Daybreak to audit software, find vulnerabilities and patch security gaps, with early results from ENISA and CERT-EU and Ukraine using it to protect civilian infrastructure.
The Texas-based dental provider says hackers accessed names, Social Security numbers, medical and insurance details, with the total reportedly possibly surpassing 455,000.
Attackers have sent over 688,000 malicious requests at 42 managed service provider servers since October 20, exploiting a bug patched in September.
Attackers injected malicious JavaScript through Optimizely's Web Experimentation platform, hitting sites including the BBC, CNN and Salesforce. The flaw is patched and the code removed.
Quarterly updates span versions 14 to 18, with a record CVE count and follow-up work required for some fixes.
Attackers reached servers holding details of thousands of users and tried stolen credentials on other machines, prompting RIKEN to take the supercomputer offline.
The FBI and Amazon Threat Intelligence say the open-source AI penetration testing framework is being used in attacks on banks, IT providers and crypto firms, with some intrusions progressing from initial targeting to privileged access in under three hours.
PUBLISHED
Canonical's new Edge archive serves Surge-generated CVE fixes before full testing, while normal Ubuntu security channels stay human-validated.
An agentic AI hacking framework called Artex touched about 10 Barracuda cloud accounts in roughly 44 hours, and the FBI warns it is spreading among Chinese threat actors.
Google's Project Zero will phase out its 90-day disclosure deadline, dropping to 60 days in 2027 and 30 days in 2028, in a shift set against AI-assisted vulnerability discovery.
Stolen Fireworks credentials let attackers push malicious Claude Code extension copies to up to 5,000 Cursor users, with Notion and Uber also hit.
Google Threat Intelligence Group reports a suspected Russian-speaking operator ran an autonomous framework that diffed code updates, built exploits and attacked victims with no human involvement.
CVE-2026-114713 lets ordinary IRC users run commands with server privileges, and CISA has added it to its exploited vulnerabilities list.