Skip to main content

Insider threat hunting

How an AI SOC hunts insider risk: compare downloads and repository clones with each user's baseline, add role context, and hand a dossier to the review board.

Use case
Updated

The short answer

Insider threat hunting looks for people misusing access they already have, such as bulk downloads from document stores or code repositories. Cotool's example Insider Risk Responder starts from Jira security issues and attaches a dossier for the review board.[1]

The problem

Information repositories such as SharePoint, Confluence, and code repositories hold data that can be collected with ordinary access and shared outside the organization.[7]

Findings about a person usually go to a review board with a dossier rather than straight to containment.[1]

How an AI SOC handles it

  1. Take the case

    A Jira security issue triggers the Insider Risk Responder.[1][2]

  2. Baseline the activity

    The agent compares the person's Google Drive downloads with their own baseline, for example a contractor at 2.4 times their normal.[1][3]

  3. Check code access

    The agent pulls GitHub audit log events and joins them with Okta sessions and device posture, as in Cotool's hunt for clone bursts from unmanaged devices.[4][8]

  4. Add context

    The agent checks the person's role and dates, such as whether a contractor's downloads came just before the engagement ended.[1]

  5. Write the dossier

    Findings go into a dossier on the alert and the Jira issue for the review board.[1][5]

Example

A hypothesis-driven hunttext
Hypothesis: a stolen GitHub token is used to clone private repositories
from an unmanaged device.

1. Pull the GitHub audit log for 14 days   (queried via API, not centralized)
2. Join Okta sessions and device posture   (Cotool logs)
3. Flag clone bursts from unmanaged devices

Detection generated: Mass repo clone from unmanaged device (T1213)

The hunt shown on Cotool's Hunt page. When a hunt returns findings, the agent backtests and ships a detection so the pattern is covered next time.[4]

What people approve

  • Insider findings go to a review board with the dossier attached.[1]
  • Response agents cannot close an alert as True Positive. When an agent confirms malicious activity, it escalates the alert and a person closes it.[6]

How Cotool handles it

  • Cotool threat models your environment, forms hypotheses about attacker behavior, and runs hunts continuously. A hunt with findings ends in a new detection.[4]

Integrations used

Sources

  1. Respond to every signal (Cotool Respond) · accessed November 16, 2026
  2. Triggering agents (Cotool docs) · accessed November 16, 2026
  3. Changelog (Cotool docs) · accessed November 16, 2026
  4. Threat hunting (Cotool Hunt) · accessed November 16, 2026
  5. Jira integration (Cotool docs) · accessed November 16, 2026
  6. Alerts (Cotool docs) · accessed November 16, 2026
  7. Data from Information Repositories, T1213 (MITRE ATT&CK) · accessed November 16, 2026
  8. Reviewing the audit log for your organization (GitHub Docs) · 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