← All playbooks
HighSupply chain · 2026-11-18

PLAID KARTE breach: stolen integration secret keys and customer data hit JINS, DHC, J.League and other Japanese brands

SockAnalyst · Security+ certified analyst.

What happened

Japanese SaaS company PLAID says an attacker exploited a vulnerability on one of its servers in the early hours of November 13, 2026 (JST), stole credentials used to connect to its systems and databases, read data out of those databases, and used the stolen credentials to take over PLAID accounts and reach the KARTE management console. PLAID detected the intrusion at 8:59 a.m. JST, disabled the stolen credentials, began shutting servers down around 2:27 p.m., put every KARTE service into maintenance mode and finished containment by about 3:23 p.m.; it says it has seen no new unauthorized access since. In its November 16 disclosure PLAID confirmed the leak of 79,249 email addresses registered as KARTE accounts, of the secret keys and access tokens customers use to connect KARTE to outside services (count still under investigation), and of part of the member data client companies imported into KARTE (for example through its linking-table feature), which can include personal information depending on what each client sent. PLAID has not named the vulnerability, and KARTE remains offline with no restart date. Client companies have since disclosed their own exposure: J.League says 2,383,073 ticketing-site records (internal user IDs, birth dates, gender, prefecture and city) and 543 Fantasy Card records leaked; JINS says up to about 18.29 million member accounts are in scope; Nikkei reports DHC disclosed about 19.11 million customer account records on November 18; and Mitsubishi UFJ Trust and Banking and other firms have also reported impact. Companies say login passwords and card numbers were not in the KARTE data. Totals overlap across brands and cannot simply be added up.

Who's affected

Every company that used KARTE, especially those that configured KARTE integrations with outside services (their secret keys and access tokens leaked), and the customers of those companies, whose member attributes may have been imported into KARTE. The 79,249 KARTE account holders (mostly client-company staff) are likely phishing targets, and end customers of the named brands should expect phishing that impersonates them.

✅ Fix / patch

  • KARTE customers: revoke and reissue every secret key, API key and access token you gave KARTE for integrations (PLAID's explicit request). Until you do, PLAID warns attackers could use them to write data into or read data out of the connected services.
  • After rotating, review audit logs of each connected service (cloud storage, data warehouse, CRM, messaging/LINE, email delivery) for use of the old KARTE credentials since November 13, 2026, and for any access from unfamiliar IPs.
  • Inventory exactly which customer fields you sent to KARTE (linking tables, event data) so you can scope notification; request the leaked-data copy PLAID is providing to clients that ask.
  • Warn staff whose email is a KARTE login and your own customers about phishing that impersonates PLAID or your brand; do not open links in unexpected 'KARTE' or account-check emails.
  • Going forward: give marketing SaaS the least-privileged, write-only or scoped tokens possible, set expiry dates, and minimize the personal fields you export to third parties.

🔎 Quick hunt check

No IOCs (IPs, domains, hashes) or CVE have been published by PLAID or affected companies as of Nov 18, 2026, so hunt on the two exposures that are confirmed. (1) Leaked integration credentials: find the service principals, API keys or IAM users you created for KARTE (names usually contain karte or plaid; otherwise use your integration inventory) and flag any use since November 13, 2026, especially from new IP addresses, new user agents or outside normal sync times; any use after you rotated them is malicious. (2) Follow-on phishing: look for inbound email mentioning KARTE or PLAID (プレイド) from senders that are not plaid.co.jp / karte.io, and for lookalike domains.

MITRE ATT&CK: T1190, T1552, T1078.004, T1530, T1199, T1566

IOCs

No IOCs (IPs, domains or hashes) published by the sources yet.

Hunt queries

Starting points built only from published indicators. Field names vary; tune before use.

```PLAID KARTE breach (Nov 13, 2026): use of KARTE integration credentials + KARTE/PLAID-themed phishing```
index=* earliest=11/13/2026:00:00:00
("karte" OR "plaid" OR "プレイド")
| eval actor=lower(coalesce(user, userName, src_user, app, client_id, ""))
| eval subj=coalesce(subject, Subject, "")
| eval sender=lower(coalesce(src_user, sender, From, ""))
| eval signal=case(
    match(actor,"karte|plaid") AND match(lower(coalesce(action, eventName, operationName, "")),"(?i)get|list|read|export|download|put|write|post|create|update|select|query"), "karte_integration_credential_used",
    match(subj,"(?i)karte|plaid|プレイド") AND isnotnull(sender) AND sender!="" AND NOT match(sender,"@(plaid\.co\.jp|karte\.io)$"), "karte_themed_email_external_sender",
    true(), null())
| where isnotnull(signal)
| eval client=coalesce(src, src_ip, sourceIPAddress, clientip)
| stats count min(_time) as first_seen max(_time) as last_seen values(client) as source_ips values(eventName) as actions values(subj) as subjects by signal, actor, sourcetype
| convert ctime(first_seen) ctime(last_seen)
| sort - count
// PLAID KARTE breach (Nov 13, 2026): use of KARTE integration credentials + KARTE/PLAID-themed phishing
let Since = datetime(2026-11-13);
union isfuzzy=true
(AADServicePrincipalSignInLogs | where TimeGenerated >= Since
 | where ServicePrincipalName has_any ("karte","plaid")
 | extend Signal = "karte_service_principal_signin", Who = ServicePrincipalName, Evidence = strcat(IPAddress, " -> ", ResourceDisplayName, " result ", ResultType)),
(AWSCloudTrail | where TimeGenerated >= Since
 | where UserIdentityUserName has_any ("karte","plaid") or UserIdentityArn has_any ("karte","plaid")
 | extend Signal = "karte_aws_credential_used", Who = coalesce(UserIdentityUserName, UserIdentityArn), Evidence = strcat(SourceIpAddress, " ", EventName, " ", UserAgent)),
(EmailEvents | where TimeGenerated >= Since
 | where (Subject has_any ("KARTE","PLAID","プレイド") or SenderDisplayName has_any ("KARTE","PLAID","プレイド"))
 | where not(SenderFromDomain in~ ("plaid.co.jp","karte.io"))
 | extend Signal = "karte_themed_email_external_sender", Who = RecipientEmailAddress, Evidence = strcat(SenderFromAddress, " | ", Subject))
| summarize Hits = count(), FirstSeen = min(TimeGenerated), LastSeen = max(TimeGenerated), Signals = make_set(Signal), Evidence = make_set(Evidence, 20) by Who
| order by Hits desc

See all hunt queries →

⏱ CISA deadline

Not in CISA KEV as of Nov 18, 2026.

Sources

Related playbooks

Get the SockAnalyst Weekly Brief free, every Thursday.

One email a week with the KEV deadlines, patches and hunt queries that matter. New playbooks still post here on the site every morning.

Subscribe free

Read the free weekly pack →