An old incident becomes a new data point
Harvard discovered unauthorized access to its Alumni Affairs and Development systems on November 18, 2025, after what the university has described as a phone-based vishing attack, meaning a caller impersonated someone to extract credentials or access over the phone rather than through a technical exploit. Harvard notified affected individuals by email on November 22, 2025, stating it had removed the unauthorized access and found no evidence it continued afterward. The university's notice also said the affected systems did not generally contain passwords, Social Security numbers, payment card data, or bank account information.
What makes this a story again a year later is that Have I Been Pwned added the associated dataset to its public breach catalog on November 3, 2026, listing roughly 851,000 unique email addresses. HIBP's own count, about 850,500, is close enough to be the same figure rounded differently, and the service is careful to note this reflects unique email addresses in the dataset, not a confirmed count of individuals whose information was accessed or compromised.
What is actually in the exposed data
According to HIBP's cataloging, the dataset includes names, phone numbers, physical addresses, birth dates, employer information, donation history, citizenship status, ethnicity, and family members' names, with individual records varying in how many of those fields they actually contain. This is precisely the kind of information that enables highly convincing social engineering rather than direct financial fraud, since an attacker armed with someone's donation history, employer, and family details can construct a phishing pretext that passes every instinctive sniff test a recipient would normally apply to an unfamiliar email or call.
In February 2026, the group ShinyHunters claimed responsibility for the breach and published a roughly 1.1GB archive it said contained more than one million records. Reporting at the time indicated some of the published information was independently verified as accurate, though Harvard has not publicly attributed the intrusion to ShinyHunters by name. A separate attack on Instructure's Canvas platform in May 2026 has also been linked to the same group, though no established connection ties that incident to the Harvard breach specifically.
Why the one-year gap matters operationally
Harvard's own incident response appears to have closed in November 2025, with the university stating it found no evidence of continued unauthorized access. But the risk to the individuals in that dataset did not close with the incident response. ShinyHunters' February 2026 publication and this week's addition to Have I Been Pwned represent two separate moments, three and twelve months after discovery respectively, when the exposed data became newly actionable for a different set of potential attackers who had no involvement in the original intrusion.
This is the structural problem with treating breach response as a bounded project with a start and end date. The individuals affected, largely alumni and donors, remain targets for phishing campaigns built on this data indefinitely, and each public re-surfacing of the dataset, whether through a dark web archive or a breach notification service, effectively resets the clock on active exploitation risk without resetting anyone's awareness that they are still exposed.
The vishing entry point deserves its own attention
We think the attack vector here is underdiscussed relative to the record count. A phone-based social engineering call succeeding against a university's development office systems points to a gap that technical controls alone will not close, since the attacker's path to access ran through a person, not a software vulnerability. Development and alumni affairs offices sit on exactly the kind of high-value personal and financial-adjacent data that makes them attractive vishing targets, yet they are rarely included in the same phishing-resistant authentication rollouts that IT departments prioritize for finance or executive staff.
Any organization running donor, alumni, or high-net-worth client relationship systems should treat those teams as a distinct high-risk population for social engineering training and phishing-resistant MFA, on the same tier as finance and executive assistants, rather than assuming general employee awareness training covers the gap. These teams exist specifically to build trust and respond quickly to requests from people claiming familiarity with the organization, which is exactly the disposition a vishing caller is trained to exploit, and that tension deserves its own tailored training rather than a slide added to the annual compliance module.
What this means for your roadmap
If your organization holds donor, alumni, or similarly rich relationship data, treat the systems that store it as a crown-jewel asset requiring phishing-resistant MFA and vishing-specific social engineering training for the staff who manage it, not just standard security awareness content. Subscribe your organization's domains to continuous breach-monitoring feeds like Have I Been Pwned's domain search rather than checking only at the time of a known incident, since this case shows meaningful new exposure events can surface many months after the original breach closed.
Finally, build a standing communication plan for re-notifying affected individuals when a dataset resurfaces publicly, even if your original incident response is long closed, since a recipient who forgot about a notice from a year ago is exactly the kind of target a phishing campaign built on this data is counting on. A short reminder email timed to a public re-surfacing event costs little and gives your affected population a fresh, accurate reason to be suspicious of unexpected donation or membership requests in the weeks that follow.



