Skip to main content

Identity compromise response

How an AI SOC responds to a compromised account: confirm from sign-ins and risk signals, scope what the account touched, then cut off access with approval.

Use case
Updated

The short answer

Identity compromise response confirms that someone else is using a valid account, finds what the account touched, and cuts off access. Cotool's example Identity Alert Responder triages Okta Identity Protection alerts, retrieves the user, and can revoke their sessions when compromise is confirmed.[1]

The problem

Adversaries use stolen credentials for initial access, persistence, and privilege escalation, and to bypass access controls on VPNs, webmail, and other remote services.[6]

Revocation isn't instant everywhere: Microsoft notes a delay of a few minutes after revokeSignInSessions before tokens are revoked.[7]

How an AI SOC handles it

  1. Confirm from sign-ins

    The agent pulls the user's recent sign-ins from Okta or Entra ID, including conditional access outcomes, and compares location, device, and network with the user's history.[2][3]

  2. Check risk signals

    Okta Identity Protection alerts and Entra ID risk detections add the identity provider's own view. Microsoft requires an Entra ID P1 or P2 license for the risk detection API.[1][2][8]

  3. Scope what it touched

    The agent lists the account's new OAuth grants in Google Workspace and the shared-vault items it viewed or copied in 1Password.[3][9][10]

  4. Cut off access with approval

    The agent proposes revoking Okta or Entra ID sessions, disabling the Entra user, or resetting the Google Workspace password, and each action waits for a person in the app or Slack.[1][3][4]

  5. Rotate exposed secrets

    The vault items the account copied are listed for their owners, so responders know which secrets to rotate.[10]

Example

Pull sign-ins, then revoke sessionshttp
GET https://graph.microsoft.com/v1.0/auditLogs/signIns?$filter=createdDateTime ge 2026-11-09T00:00:00Z and createdDateTime le 2026-11-16T23:59:59Z

POST https://graph.microsoft.com/v1.0/users/{id | userPrincipalName}/revokeSignInSessions

The two Microsoft Graph calls in this flow: list a week of sign-ins, then revoke the user's refresh tokens and browser session cookies once approved.[7][11]

What people approve

  • Session revocation, disabling a user, and password resets are state-changing actions that can require approval in the app or in Slack.[4]
  • Response agents cannot close an alert as True Positive. When an agent confirms malicious activity, it escalates the alert and a person closes it.[5]

How Cotool handles it

  • Cotool added response actions to reset Google Workspace passwords, disable Microsoft Entra users, and revoke their sessions in release 0.76.0.[3]
  • People can direct the responder from a Slack thread, for example by replying "Revoke the sessions".[1]

Integrations used

Sources

  1. Respond to every signal (Cotool Respond) · accessed November 16, 2026
  2. Microsoft Graph integration (Cotool docs) · accessed November 16, 2026
  3. Changelog (Cotool docs) · accessed November 16, 2026
  4. AI SOC (Cotool) · accessed November 16, 2026
  5. Alerts (Cotool docs) · accessed November 16, 2026
  6. Valid Accounts, T1078 (MITRE ATT&CK) · accessed November 16, 2026
  7. user: revokeSignInSessions (Microsoft Graph) · accessed November 16, 2026
  8. List riskDetections (Microsoft Graph) · accessed November 16, 2026
  9. Manage OAuth token audit activity (Google Workspace Admin SDK) · accessed November 16, 2026
  10. 1Password Events API (1Password Developer) · accessed November 16, 2026
  11. List signIns (Microsoft Graph) · accessed November 16, 2026

Related use cases

All use cases

Related reading

Attackers are scaling with tokens

Cotool helps defenders operate at machine speed. See how security teams are scaling Detection & Response.

Book a demo