← 返回列表

How to Investigate GitHub PAT Compromise: Lessons From a Multi-Organization Campaign

单站 Wiz Blog 作者 Eden Abergil 发布 Thu, 13 Aug 2026 15:04:28 GMT 入库 2026-11-18T05:29:58

Organizations increasingly rely on GitHub to develop and store proprietary source code, internal documentation, and other software assets. This makes GitHub Personal Access Tokens (PATs) an attractive target for attackers, as a compromised token can provide access to private repositories and expose secrets such as cloud credentials, API keys, and private keys that may enable further compromise.

Recently, the Wiz Customer Incident Response Team (CIRT) investigated a coordinated campaign in which compromised GitHub PATs were used to conduct repository reconnaissance and mass repository exfiltration across multiple organizations. Active from mid-May through early June 2026, the campaign progressed through several distinct stages, from reconnaissance and access validation to large-scale repository cloning and follow-on attempts to leverage exfiltrated credentials.

In this blog, we walk through a practical approach to investigating GitHub PAT compromise. Using a real-world campaign investigated by Wiz CIRT as an example (Wiz customers can see the Threat Intel Center advisory for campaign details and IOCs), we demonstrate how to reconstruct the attack timeline, identify the full blast radius, and hunt for follow-on activity.

For most impacted organizations, the observable indication of the campaign was a significant increase in git.clone activity on June 1, as seen in GitHub Audit Logs. Between 09:14 and 14:55 UTC, the actor used 102 AWS IP addresses located in the ca-central-1 region to clone up to thousands of repositories per organization. The activity used the git/2.43.0 user agent and was highly parallelized, indicating automated tooling.

The cloning operations were performed using valid PATs belonging to employees of the affected organizations.

The first priority is to identify the PAT used in the cloning activity and prevent any unauthorized repository access using the compromised PAT.

Containment options depend on the type of PAT involved and the organization’s identity configuration:

Because organization owners may be unable to independently verify the revocation of a PAT or rule out the compromise of additional tokens associated with the user, the user should also be temporarily removed from the organization while the investigation continues.

Because a compromised PAT may indicate broader compromise of the user’s endpoint or other credential stores, responders should also consider restricting the user’s access to other sensitive systems and isolating potentially affected endpoints until the source and scope of the compromise have been determined.

The next step is to expand the investigation by reviewing GitHub activity associated with the compromised PATs.

Responders should first establish the user's normal activity patterns, using GitHub audit logs, including typical source IP addresses, geolocations, user agents, access times, and repository activity, and use this baseline to identify anomalous events associated with the compromised PAT or user.

Responders should also review the user’s personal GitHub security log, as it may contain relevant personal account events that are not recorded in the organization audit log. In EMU environments, equivalent user-level security events are available in the enterprise audit log. In non-EMU environments, organization owners generally cannot access a personal user’s security log directly, so the affected user may need to review or export this information for the investigation.

In parallel, pivot on any known indicators of compromise (IOCs), such as source IP addresses and user agents, to identify additional related activity that may not be associated with the compromised user.

In confirmed compromise scenarios, responders should also consider engaging GitHub Support and providing relevant details about the incident. Depending on the circumstances, GitHub may be able to provide additional logs that are not directly available to the organization, potentially offering further visibility into the compromise.

Wiz customers can use the following Wiz Cloud Events query to retrieve activity conducted by a compromised PAT in order to identify anomalies. Additionally, use the Activity tab in the GitHub user’s drawer to see additional queries that can help with baselining like “Source ASO for Resource” and “Events by Resource” over time.

In this investigation, this approach uncovered two earlier stages that preceded the mass cloning event.

Prior to the mass clone, on May 15, the threat actor used the GitHub API to query the /repositories/{id}/readme endpoint from a single AWS IP address (13.221.167.217). Based on the AWS IP address documentation, this IP address is associated with the EC2 services and located in us-east-1. The requests originated from a standard Chrome browser user agent (Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Safari/537.36).

These requests likely allowed the actor to collect information regarding accessible repositories and identify targets before performing any cloning activity.

One important observation is that these requests generate api.request events, which are available through GitHub Enterprise Log Streaming but are not included in standard GitHub audit logs. Organizations without Enterprise Log Streaming therefore do not have visibility into this reconnaissance activity.

Additional related activity occurred between May 29 and May 31. The same compromised PATs were used to perform a limited number of git.clone operations from 107.174.201.183 (HostPapa, US) using the git/2.25.1 user agent.

Unlike the June 1 activity, only a small subset of repositories was cloned. These cloning operations began within seconds of one another (approximately 05:42-05:44 UTC) across several different victims, suggesting automated tooling.

This activity appears to represent a validation phase.

At this stage, the investigation had reconstructed the following three phases of the campaign:

With the attack timeline reconstructed, the next objective was to determine what the attacker had actually obtained and assess the resulting blast radius.

After expanding the timeline, the next step is determining how the compromised token was exposed. Identifying the source of the leaked PAT may reveal additional leaked credentials.

Common sources include endpoint compromise or plaintext secrets in cloud resources (e.g., hardcoded in code, configuration files, or environment variables).

Use the following Wiz Graph query to search for Secret Findings associated with the compromised hashed_token.

In this investigation, across all impacted organizations, the initial access vector was not identified. Our investigation did not identify the compromised tokens in code or cloud resources, and we found no evidence indicating that they were exposed through those environments.

The method by which the actor obtained valid PATs across multiple unrelated organizations therefore remains an open question. Given that the investigation did not identify a code or cloud-based source of exposure, and based on Wiz CIRT’s experience with credential theft, endpoint compromise is considered a possible initial access vector, although this could not be confirmed.

The next objective is to determine what the attacker obtained and assess the resulting blast radius.

Repository exfiltration should be treated as more than source code theft. Repositories frequently contain cloud credentials, SaaS tokens, internal documentation, private keys, and sensitive data that can significantly increase the impact of a compromise.

Investigators should account for GitHub’s audit-log retention limitations when assessing repository access. In GitHub Enterprise Cloud, Git events such as git.clone, git.fetch, and git.push are retained by GitHub for only seven days. If these events were not already streamed to an external system, such as Wiz, activity outside this retention window may no longer be available, potentially limiting the ability to establish the full scope of repository exfiltration.

To understand the full scope of the breach, use the GitHub audit logs to identify every repository the actor cloned. This serves as the foundation for the next steps - identifying any secrets and sensitive data exposed within those repositories.

Use the following Wiz Cloud Events query with any relevant IOCs (IPs, user agents, hashed tokens) to list all affected repositories. The search results can be extracted to a CSV to be used in the next steps.

Identify Exposed Secrets

Review the affected repositories for secrets that could provide access to additional systems, including cloud provider credentials, SaaS tokens, private keys, database credentials, CI/CD secrets, and internal service credentials.

Use the following Wiz Cloud Events query with any relevant IOCs (IPs, user agents, hashed tokens) to list all affected repositories. The search results can be exported to a CSV to be used in the next steps.

https://app.wiz.io/findings/secret-findings#~(streamCols~(~'event~'principal~'principalIp~'resource)~filters~(severity~(notEquals~(~'INFORMATIONAL))~status~(equals~(~'OPEN))~codeRepository~(equals~(~'<RESOURCE_ID>~'<RESOURCE_ID>))))
To prioritize remediation:

Any valid secret present within an exfiltrated repository should be treated as potentially compromised.

Identify Sensitive Data

Repositories may also contain regulated or business-sensitive information including PII, PHI, business confidential information, and financial information. Identifying such information helps determine the organization’s full exposure and assess whether regulatory notification or disclosure obligations may apply.

Use Wiz Data Findings with the repository Resource IDs from the CSV to determine which data may exist in these repositories:

https://app.wiz.io/findings/data-findings/classification-and-compound#~(filters~(status~(equals~(~'OPEN))~vcsRepository~(equals~(~'<RESOURCE_ID>~'<RESOURCE_ID>)))~streamCols~(~'event~'principal~'principalIp~'resource))

Any credential present within an exfiltrated repository should be treated as potentially compromised.

Before rotating credentials, document each rotated credential, including the credential identifier, affected service and rotation time, as this information will be required during the next phase of the investigation.

Once the inventory has been established, rotate all exposed credentials as quickly as possible, including cloud provider credentials, SaaS API keys, GitHub tokens, private keys, certificates, database credentials, CI/CD credentials, and internal service accounts.

After collecting the list of exposed credentials, search for any unauthorized activity performed using those secrets prior to rotation.

This can help determine whether the actor moved beyond repository access into cloud or other environments. Hunt across all relevant services - cloud provider credentials (e.g., AWS, GCP, Azure), SaaS platform tokens (e.g., GitHub, GitLab), internal service credentials etc.

Example Wiz Cloud Events query that searches for activity associated with the leaked AWS access key and summarizes the results by user agent and IP address to help identify anomalous activity.

GitHub PAT compromise can escalate quickly - from initial reconnaissance to mass repository exfiltration and subsequent credential abuse across cloud environments. The investigation playbook outlined above - contain, expand the timeline, identify the leak source, assess blast radius, rotate credentials, and hunt for unauthorized use - provides a structured approach for responders to work through methodically under pressure.

A key challenge in investigating GitHub PAT compromise is that the evidence spans multiple disconnected systems - GitHub audit logs, cloud provider control planes, and code repositories. As demonstrated throughout this post, Wiz unifies these signals into a single investigation surface - from querying GitHub audit log activity and identifying exposed secrets across exfiltrated repositories, to hunting for unauthorized use of leaked credentials in cloud environments. This unified visibility reduces investigation time and minimizes the risk of missing follow-on activity that crosses from GitHub into cloud environments.

To improve detection and response for these types of attacks, Wiz recommends that organizations stream GitHub audit logs to Wiz, preferably enterprise-level audit logs, alongside cloud provider logs, to enable cross-platform threat correlation.

Correlating GitHub activity with cloud control-plane events makes it possible to trace an attacker's path from a compromised PAT through to unauthorized cloud access across the full attack chain.

Wiz Defend includes Threat Detection Rules (TDRs) that detect suspicious and malicious activity across GitHub environments. Some of these rules reflect techniques observed in this campaign, while others provide broader coverage for GitHub account, credential, and repository compromise. The following rules help organizations detect activity across different stages of GitHub-based attacks:

Rule Name Rule ID
Mass Clone Operations Of Private Repositories Performed By User cer-github-data-massPrivateRepoCloneOperationsByUser
Mass Clone Operations Of Private Repositories Performed By Bot cer-github-data-massPrivateRepoCloneOperationsByBot
Private Repository Cloned From Low Reputation IP cer-github-data-privateRepositoryClonedFromLowReputationIP
Personal Access Token Used From Unusual ASO cer-github-control-personalAccessTokenUsedFromUnusualASO
Unusual user access to Github through third-party VPN cer-github-control-unusualUserAccessThroughThirdPartyVPN
Automated High Volume Activity From Non Bot User cer-github-control-automatedHighVolumeActivityFromNonBotUser
Personal Access Token Used After Long Period of Inactivity cer-github-identity-personalAccessTokenUsedAfterLongPeriodOfInactivity
Private/Internal Repository Enumerated By User cer-github-data-repoEnumeratedByUser
Private/Internal Repository Access From Unusual ASO cer-github-control-repoAccessedFromUnusualASO
Workflow Log Deleted cer-github-control-workflowLogDeleted

If you experience a GitHub-related breach or any cloud security incident, contact Wiz CIRT for incident response support.

IOC Type IOC Description
User Agent Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Safari/537.36 User-agent used in the first phase for reconnaissance
User Agent git/2.25.1 User-agent used in the second phase for probing
User Agent git/2.43.0 User-agent used in the third phase for exfiltration

IP First seen Last Seen Country ASO
13.221.167.217 2026-05-15 2026-05-15 United States Amazon.com, Inc.
107.174.201.183 2026-05-29 05:42:24.849 2026-05-31 11:30:09.136 United States HostPapa
16.52.86.223 2026-06-01 09:14:10.513 2026-06-01 09:14:24.822 Canada Amazon.com, Inc.
15.223.74.214 2026-06-01 09:14:11.266 2026-06-01 14:53:04.456 Canada Amazon.com, Inc.
35.183.65.4 2026-06-01 09:14:11.366 2026-06-01 09:20:57.986 Canada Amazon.com, Inc.
15.157.62.127 2026-06-01 09:14:11.408 2026-06-01 09:34:53.299 Canada Amazon.com, Inc.
35.183.112.105 2026-06-01 09:14:11.481 2026-06-01 09:32:06.972 Canada Amazon.com, Inc.
99.79.193.97 2026-06-01 09:14:11.596 2026-06-01 09:19:19.136 Canada Amazon.com, Inc.
52.60.229.238 2026-06-01 09:14:11.842 2026-06-01 09:20:05.054 Canada Amazon.com, Inc.
35.182.157.210 2026-06-01 09:14:12.874 2026-06-01 09:53:52.565 Canada Amazon.com, Inc.
15.223.204.99 2026-06-01 09:14:14.280 2026-06-01 09:58:47.968 Canada Amazon.com, Inc.
3.96.138.214 2026-06-01 09:17:50.670 2026-06-01 09:24:45.462 Canada Amazon.com, Inc.
15.223.67.252 2026-06-01 10:52:13.306 2026-06-01 11:01:50.331 Canada Amazon.com, Inc.
3.99.161.201 2026-06-01 10:52:13.411 2026-06-01 10:52:49.919 Canada Amazon.com, Inc.
3.96.210.9 2026-06-01 10:52:14.606 2026-06-01 11:01:44.984 Canada Amazon.com, Inc.
35.183.178.163 2026-06-01 10:52:15.243 2026-06-01 11:01:52.266 Canada Amazon.com, Inc.
35.183.235.214 2026-06-01 10:52:15.816 2026-06-01 11:01:47.971 Canada Amazon.com, Inc.
99.79.75.249 2026-06-01 10:52:18.980 2026-06-01 11:01:40.593 Canada Amazon.com, Inc.
3.96.199.215 2026-06-01 12:02:59.457 2026-06-01 12:09:09.510 Canada Amazon.com, Inc.
3.99.171.218 2026-06-01 12:02:59.831 2026-06-01 12:07:03.918 Canada Amazon.com, Inc.
15.157.65.4 2026-06-01 12:03:01.092 2026-06-01 12:06:53.165 Canada Amazon.com, Inc.
15.223.188.87 2026-06-01 12:03:03.012 2026-06-01 12:06:39.638 Canada Amazon.com, Inc.
15.223.65.175 2026-06-01 12:03:04.466 2026-06-01 12:44:41.613 Canada Amazon.com, Inc.
3.96.208.217 2026-06-01 12:03:04.558 2026-06-01 12:44:47.860 Canada Amazon.com, Inc.
15.223.63.237 2026-06-01 12:03:05.363 2026-06-01 12:45:24.660 Canada Amazon.com, Inc.
15.222.64.207 2026-06-01 12:03:09.699 2026-06-01 12:06:23.093 Canada Amazon.com, Inc.
15.157.63.171 2026-06-01 12:03:27.821 2026-06-01 12:46:34.449 Canada Amazon.com, Inc.
35.182.213.73 2026-06-01 12:03:40.439 2026-06-01 12:45:53.941 Canada Amazon.com, Inc.
16.52.82.13 2026-06-01 12:03:41.508 2026-06-01 12:06:59.332 Canada Amazon.com, Inc.
16.54.253.67 2026-06-01 12:03:56.851 2026-06-01 12:07:14.433 Canada Amazon.com, Inc.
15.157.68.0 2026-06-01 12:03:57.597 2026-06-01 12:45:50.974 Canada Amazon.com, Inc.
3.96.173.139 2026-06-01 12:04:06.765 2026-06-01 12:07:17.790 Canada Amazon.com, Inc.
99.79.72.217 2026-06-01 12:04:12.833 2026-06-01 12:38:45.410 Canada Amazon.com, Inc.
35.182.176.228 2026-06-01 12:05:17.864 2026-06-01 12:44:51.814 Canada Amazon.com, Inc.
15.223.186.60 2026-06-01 12:05:44.580 2026-06-01 12:44:41.614 Canada Amazon.com, Inc.
16.52.76.162 2026-06-01 12:05:50.860 2026-06-01 12:44:36.069 Canada Amazon.com, Inc.
15.223.2.184 2026-06-01 12:05:54.440 2026-06-01 12:09:18.594 Canada Amazon.com, Inc.
15.222.60.246 2026-06-01 12:05:55.594 2026-06-01 12:45:38.315 Canada Amazon.com, Inc.
16.52.169.32 2026-06-01 12:06:00.787 2026-06-01 12:46:40.296 Canada Amazon.com, Inc.
99.79.128.151 2026-06-01 12:06:09.722 2026-06-01 12:09:15.024 Canada Amazon.com, Inc.
15.223.200.124 2026-06-01 12:06:11.964 2026-06-01 12:44:48.575 Canada Amazon.com, Inc.
35.183.69.26 2026-06-01 12:06:16.386 2026-06-01 12:45:44.507 Canada Amazon.com, Inc.
35.183.111.110 2026-06-01 12:06:16.518 2026-06-01 12:45:31.283 Canada Amazon.com, Inc.
35.183.42.203 2026-06-01 12:06:19.816 2026-06-01 12:46:12.412 Canada Amazon.com, Inc.
99.79.48.16 2026-06-01 12:06:22.201 2026-06-01 12:45:49.431 Canada Amazon.com, Inc.
35.183.26.57 2026-06-01 12:06:27.093 2026-06-01 12:47:19.765 Canada Amazon.com, Inc.
35.182.17.157 2026-06-01 12:06:29.642 2026-06-01 12:44:43.955 Canada Amazon.com, Inc.
99.79.128.234 2026-06-01 12:06:32.076 2026-06-01 12:45:53.049 Canada Amazon.com, Inc.
99.79.33.200 2026-06-01 12:06:32.419 2026-06-01 12:09:42.988 Canada Amazon.com, Inc.
16.54.241.145 2026-06-01 12:06:35.181 2026-06-01 12:09:56.545 Canada Amazon.com, Inc.
16.52.50.245 2026-06-01 12:06:35.515 2026-06-01 12:10:22.549 Canada Amazon.com, Inc.
35.183.121.31 2026-06-01 12:06:43.566 2026-06-01 12:10:20.187 Canada Amazon.com, Inc.
35.183.34.62 2026-06-01 12:06:47.354 2026-06-01 12:09:49.736 Canada Amazon.com, Inc.
3.97.11.39 2026-06-01 12:06:53.506 2026-06-01 12:10:06.785 Canada Amazon.com, Inc.
35.183.243.11 2026-06-01 12:06:55.090 2026-06-01 12:07:24.377 Canada Amazon.com, Inc.
16.52.40.180 2026-06-01 12:06:56.979 2026-06-01 12:41:45.329 Canada Amazon.com, Inc.
99.79.52.86 2026-06-01 12:36:39.828 2026-06-01 12:47:00.050 Canada Amazon.com, Inc.
35.183.15.41 2026-06-01 12:37:16.207 2026-06-01 12:44:49.497 Canada Amazon.com, Inc.
35.182.79.86 2026-06-01 12:42:12.034 2026-06-01 12:44:21.617 Canada Amazon.com, Inc.
35.182.13.97 2026-06-01 12:42:12.905 2026-06-01 12:44:41.530 Canada Amazon.com, Inc.
99.79.64.18 2026-06-01 12:42:26.297 2026-06-01 12:44:24.993 Canada Amazon.com, Inc.
35.183.64.193 2026-06-01 12:42:37.464 2026-06-01 12:44:54.702 Canada Amazon.com, Inc.
15.223.206.124 2026-06-01 12:42:56.064 2026-06-01 12:45:12.630 Canada Amazon.com, Inc.
35.182.83.50 2026-06-01 12:44:16.002 2026-06-01 12:46:18.999 Canada Amazon.com, Inc.
35.183.106.137 2026-06-01 14:44:49.695 2026-06-01 14:49:32.812 Canada Amazon.com, Inc.
35.183.129.254 2026-06-01 14:44:56.208 2026-06-01 14:53:51.857 Canada Amazon.com, Inc.
35.183.236.91 2026-06-01 14:45:00.543 2026-06-01 14:54:23.621 Canada Amazon.com, Inc.
15.223.207.112 2026-06-01 14:45:00.716 2026-06-01 14:55:20.284 Canada Amazon.com, Inc.
35.183.20.241 2026-06-01 14:45:03.207 2026-06-01 14:55:18.817 Canada Amazon.com, Inc.
3.96.210.47 2026-06-01 14:45:12.227 2026-06-01 14:54:49.082 Canada Amazon.com, Inc.
3.96.200.150 2026-06-01 14:45:13.020 2026-06-01 14:54:00.665 Canada Amazon.com, Inc.
16.52.37.91 2026-06-01 14:45:30.295 2026-06-01 14:55:24.719 Canada Amazon.com, Inc.
3.96.173.8 2026-06-01 14:45:33.364 2026-06-01 14:55:18.276 Canada Amazon.com, Inc.
16.52.166.75 2026-06-01 14:45:37.790 2026-06-01 14:54:31.400 Canada Amazon.com, Inc.
16.54.243.226 2026-06-01 14:45:38.280 2026-06-01 14:54:50.596 Canada Amazon.com, Inc.
3.96.141.125 2026-06-01 14:45:39.948 2026-06-01 14:54:55.206 Canada Amazon.com, Inc.
35.183.136.221 2026-06-01 14:45:46.606 2026-06-01 14:54:47.652 Canada Amazon.com, Inc.
3.96.193.166 2026-06-01 14:45:47.597 2026-06-01 14:52:19.372 Canada Amazon.com, Inc.
3.96.142.210 2026-06-01 14:45:51.680 2026-06-01 14:55:08.840 Canada Amazon.com, Inc.
15.223.121.133 2026-06-01 14:45:52.442 2026-06-01 14:55:24.328 Canada Amazon.com, Inc.
15.222.61.83 2026-06-01 14:45:52.965 2026-06-01 14:54:51.626 Canada Amazon.com, Inc.
35.182.251.252 2026-06-01 14:45:53.632 2026-06-01 14:54:59.277 Canada Amazon.com, Inc.
15.223.74.59 2026-06-01 14:46:01.670 2026-06-01 14:54:59.019 Canada Amazon.com, Inc.
3.96.201.161 2026-06-01 14:46:07.577 2026-06-01 14:55:07.408 Canada Amazon.com, Inc.
15.157.70.253 2026-06-01 14:46:20.220 2026-06-01 14:55:28.326 Canada Amazon.com, Inc.
99.79.189.128 2026-06-01 14:46:26.235 2026-06-01 14:52:13.037 Canada Amazon.com, Inc.
3.96.218.171 2026-06-01 14:46:30.564 2026-06-01 14:52:10.186 Canada Amazon.com, Inc.
3.96.127.99 2026-06-01 14:46:30.608 2026-06-01 14:54:05.803 Canada Amazon.com, Inc.
16.52.128.252 2026-06-01 14:46:46.602 2026-06-01 14:52:31.118 Canada Amazon.com, Inc.
35.182.78.22 2026-06-01 14:46:51.665 2026-06-01 14:52:53.771 Canada Amazon.com, Inc.
35.183.126.121 2026-06-01 14:46:53.803 2026-06-01 14:52:42.374 Canada Amazon.com, Inc.
3.96.187.238 2026-06-01 14:47:13.647 2026-06-01 14:50:25.229 Canada Amazon.com, Inc.
16.52.147.129 2026-06-01 14:47:39.516 2026-06-01 14:53:53.970 Canada Amazon.com, Inc.
16.54.241.240 2026-06-01 14:47:39.737 2026-06-01 14:53:42.018 Canada Amazon.com, Inc.
99.79.32.153 2026-06-01 14:47:48.623 2026-06-01 14:53:23.101 Canada Amazon.com, Inc.
3.96.149.105 2026-06-01 14:47:54.939 2026-06-01 14:54:07.427 Canada Amazon.com, Inc.
15.223.199.223 2026-06-01 14:47:59.453 2026-06-01 14:53:57.055 Canada Amazon.com, Inc.
3.96.197.212 2026-06-01 14:48:06.822 2026-06-01 14:54:12.400 Canada Amazon.com, Inc.
99.79.161.206 2026-06-01 14:48:21.451 2026-06-01 14:54:22.073 Canada Amazon.com, Inc.
35.182.104.189 2026-06-01 14:48:32.297 2026-06-01 14:54:38.075 Canada Amazon.com, Inc.
35.183.210.27 2026-06-01 14:48:55.158 2026-06-01 14:54:59.545 Canada Amazon.com, Inc.
15.157.69.80 2026-06-01 14:49:38.914 2026-06-01 14:52:20.725 Canada Amazon.com, Inc.
16.52.147.249 2026-06-01 14:49:49.858 2026-06-01 14:52:35.094 Canada Amazon.com, Inc.
15.222.60.27 2026-06-01 14:49:50.282 2026-06-01 14:52:44.801 Canada Amazon.com, Inc.
原文链接:https://www.wiz.io/blog/investigating-github-pat-compromise