OneNym.idDiscuss deployment
English

INCIDENT RESPONSE / CLOUD / SAAS / AI

Cloud, SaaS and AI
security incidents, from first response to prevention

Hijacked accounts, leaked API keys, data exposed through generative AI. We stop the damage while preserving evidence, identify the cause and put controls in place so it does not happen again.

See all three situations →

FIRST STEPS

What to do first if you suspect an incident

DO

Do

  • Change the password of any suspicious account and sign out all of its sessions
  • If API keys or access tokens may be involved, revoke and reissue them
  • Write down when and how you noticed, and keep screenshots
  • If you have cyber insurance, notify your insurer
DON'T

Don't

  • Delete or reset the affected accounts or mailboxes
  • Wipe computers or clean up and delete logs
  • Announce the incident widely before the facts are known

Many cloud services keep logs for only a short time (Microsoft 365 sign-in logs, for example, are kept 7 to 30 days depending on the license), so preserve them as early as possible.

Request urgent help

TYPICAL INCIDENTS

Incidents we handle

01

Hijacked email accounts

Invoice fraud with changed bank details, inbox forwarding rules and suspicious mail sent to partners.

02

Malicious app consents

An app or integration granted consent through a deceptive prompt keeps reading mail, files or business data.

03

Leaked or misused API keys

Keys exposed in repositories or configuration files, and usage charges nobody recognizes.

04

Data exposed through generative AI

Confidential or personal data entered into AI, and AI tools or connectors with more access to business systems than needed.

05

Unintended AI agent actions

Prompt injection, or unauthorized actions through AI agents, coding assistants or third-party integrations such as MCP servers.

06

Stolen administrator access

Attackers staying in by adding administrators, MFA methods or mail flow rules.

SERVICES

Emergency response and readiness

The same team handles incidents when they happen and prepares you so you know what to do when one does.

Emergency incident response

STEP 1 · 24–72 HOURSFirst response
Preserve evidence, close the way in and remove any access the attacker still holds. Includes a first-response report.
STEP 2 · 1–3 WEEKSRoot-cause investigation
Rebuild the timeline and identify what data was accessed or taken, including through AI integrations. Includes a report for management, auditors and insurers.
STEP 3 · 2–4 WEEKSPrevention
Strengthen multi-factor authentication and access controls, restrict app integrations, set up log retention and alerts, and define AI usage rules.

Incident readiness package

STEP 1 · 1–2 WEEKSSecurity assessment
Review the settings of the cloud services you use and how AI is used, and report a prioritized list of improvements.
STEP 2 · 2–4 WEEKSResponse plan
Prepare a response playbook, contact tree and report templates, then run a tabletop exercise on a simulated incident.
STEP 3 · MONTHLYOngoing support
An emergency contact, monthly configuration checks, an annual exercise and staff training.

COVERAGE

Coverage

These are examples. Ask us about services or combinations not listed here.

Cloud and identity
Sign-ins, multi-factor authentication, admin rights and audit logs in Microsoft 365, Google Workspace, Entra ID, Okta and similar
Generative AI and agents
Admin settings and usage rules for ChatGPT, Claude, Microsoft Copilot, Gemini and similar, connectors to business systems, API keys, AI agents, coding assistants and MCP servers
SaaS and business apps
Sharing settings, third-party integrations and account management in Slack, Salesforce, kintone, Box and similar
Development and cloud infrastructure
Access keys, repository visibility and deployment permissions in GitHub, AWS, Azure, Google Cloud and similar
Reporting
Management reports, the facts needed for a report to Japan's Personal Information Protection Commission, draft notices to customers and partners, and material for insurers and auditors

REPORTING DEADLINES / JAPAN

Personal data incidents have reporting deadlines

Under Japan's Act on the Protection of Personal Information, a leak caused by unauthorized access or other malicious intent (even for one person), a leak of sensitive personal data, a leak that risks financial harm, or a leak affecting more than 1,000 people must be reported to the Personal Information Protection Commission and to the people affected, including when a leak is only suspected.

Preliminary report
Promptly after becoming aware, generally within 3 to 5 days
Final report
Within 30 days, or within 60 days for leaks caused by malicious intent
Notice to individuals
Promptly, as the situation allows

Confirm legal questions, such as whether you must report, with your legal counsel. OneNym.id investigates and organizes the facts the report needs.

ROLES

Who does what

OneNym.id
Investigation, containment and configuration on cloud services, SaaS and AI services, reports, playbooks and training
Specialists
Legal decisions with your counsel, court-grade forensics of computers and servers with a specialist firm, and insurance claims with your insurer.

CONTACT

Tell us what happened

Share what you can: how and when you noticed, which accounts or services may be affected, what you have already done, and whether personal data may be involved. For urgent cases we prioritize a same-day meeting.

Discuss incident response