Skip to navigation

Threat reportData BreachTL-2026-3486

Meta GitHub Repositories Breached via Stolen Employee Credentials; Source Code and Personal Data of ~1,500 People Stolen

mediumACTIVE

Attackers used stolen employee credentials to access Meta's GitHub repositories and took product source code (about 99% of the stolen data) plus a small amount of personal information (about 1%) affecting roughly 1,500 people, mostly U.S.-based Meta employees. Meta detected the intrusion several weeks before it began notifying affected individuals on 2026-10-13; the attackers have not been identified.

Key facts for Meta GitHub breach

IdentifiersTL-2026-3486
SeverityMEDIUM
First disclosed
Exploitation statusNo known exploitation
Affected productsMeta Meta GitHub-hosted source code repositories (specific repositories and products not disclosed)
Weaknesses (CWE)CWE-522, CWE-312
MITRE ATT&CK techniquesT1078 Valid Accounts, T1213.003 Code Repositories, T1552.001 Credentials In Files
MotivationUNKNOWN
Target sectorstechnology, social-media
Target regionsNorth America
Detection rules9 (1 shown free)
Sources8 cited sources

Detection coverage for Meta GitHub breach

As of 2026-10-14, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-3486 across Splunk SPL, Microsoft KQL and Sigma, covering 8 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.

Free detection: Meta GitHub Breach (TL-2026-3486) — New SSH/Deploy Key or Access Token Created on GitHub Account

Detects creation of SSH keys, deploy keys, OAuth authorizations or personal access token grants. After using stolen employee credentials (T1078) an intruder can mint persistent credentials (T1098.001, T1528) that survive a password reset. Meta revoked and reissued credentials and tokens in response.

  • Languages: Splunk SPL, Microsoft KQL, Sigma
  • ATT&CK: T1098.001, T1528, T1078
  • Data sources: github:audit_log
  • Severity: high
  • Confidence: medium

Splunk SPL

index=github sourcetype="github:cloud:audit" action IN ("public_key.create","deploy_key.create","oauth_access.create","oauth_authorization.create","personal_access_token.access_granted") earliest=-24h
| stats count as events, min(_time) as first_event, values(action) as actions, values(repo) as repos by actor, actor_ip, programmatic_access_type
| convert ctime(first_event)
| table actor, actor_ip, programmatic_access_type, actions, repos, events, first_event

Microsoft KQL

GitHubAuditLogPolling_CL
| where TimeGenerated > ago(24h)
| where action_s in ("public_key.create", "deploy_key.create", "oauth_access.create", "oauth_authorization.create", "personal_access_token.access_granted")
| summarize events = count(), first_event = min(TimeGenerated), actions = make_set(action_s), repos = make_set(repo_s) by actor_s, actor_ip_s
| project actor_s, actor_ip_s, actions, repos, events, first_event

Sigma

title: Meta GitHub Breach TL-2026-3486 - New SSH Key Deploy Key or Access Token on GitHub
id: 7c1f3a52-4b6e-4d1a-9c2b-3e5a1d0f8a02
status: experimental
description: New SSH key, deploy key, OAuth authorization or personal access token grant in the GitHub audit log, a persistence step after valid-account compromise.
references:
    - https://www.securityweek.com/meta-says-hackers-took-code-personal-data-of-1500-in-github-breach/
author: AII-Detector — ThreadLinqs Intelligence
date: 2026/10/14
tags:
    - attack.persistence
    - attack.credential_access
    - attack.initial_access
    - attack.t1098.001
    - attack.t1528
    - attack.t1078
    - threat.tl-2026-3486
logsource:
    product: github
    service: audit
detection:
    selection:
        action:
            - public_key.create
            - deploy_key.create
            - oauth_access.create
            - oauth_authorization.create
            - personal_access_token.access_granted
    condition: selection
falsepositives:
    - Developers registering new workstations or tokens
    - Approved automation onboarding
level: medium

Known false positives:

  • Developers adding SSH keys for new laptops
  • Legitimate integrations and CI authorizing OAuth apps or deploy keys

Tuning notes: Alert with higher priority when the source IP is new for the actor (join with SPL-001 baseline) or occurs within hours of a sign-in from a new country; exclude approved integration accounts.

8 more detection rules (Splunk SPL, Microsoft KQL, Sigma) · unlock with TL Creds. Sign up free to unlock the detection rules or query them with the Threadlinqs MCP server (Purple)
  • Meta GitHub Breach (TL-2026-3486) — Entra ID Sign-in to GitHub from Unfamiliar IP or Flagged Risky · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — GitHub SSO Credential Stuffing Followed by Successful Sign-in · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — Private Repository Made Public or Transferred · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — Mass Repository Clone via Single Access Token · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — GitHub Two-Factor Authentication Disabled or Requirement Removed · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — GitHub IP Allow List Disabled or Entry Removed · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — Bulk Repository Clone/Download from Previously Unseen Source IP · unlock with TL Creds
  • Meta GitHub Breach (TL-2026-3486) — Secrets Committed to Repositories / Push Protection Bypassed · unlock with TL Creds

Indicators of compromise for Meta GitHub breach

As of 2026-10-14, Threadlinqs Intelligence tracks 8 indicators of compromise for TL-2026-3486.

8 indicators of compromise · unlock with TL Creds. Compare plans

How Meta GitHub breach works

Meta disclosed on 2026-10-13 (first reported by SecurityWeek) that unidentified attackers accessed some of its GitHub repositories using stolen employee credentials. Meta's security team detected the intrusion a few weeks before the disclosure (the exact date and intrusion start were not published), cut off the attackers' access, and engaged GitHub and law enforcement. Per Meta, the incident was not caused by a vulnerability in any Meta product and Meta's internal systems were not breached; only the GitHub repositories were affected. One syndicated account (International Cyber Digest) states the credentials belonged to a single employee; Meta has not said how the credentials were stolen or whether multi-factor authentication was in place.

About 99% of the stolen material was business data, i.e. software source code related to Meta products, and about 1% was personal information concerning approximately 1,500 individuals, mostly U.S.-based Meta employees. Meta stated that it mistakenly placed personal information in the repositories despite company rules prohibiting storing it on GitHub, and that a significant portion of the apparent personal records was synthetic or anonymized test data. The specific personal data types, the number of repositories accessed, and which products' code was taken were not disclosed.

Meta's investigation with GitHub found no indication that the attackers were targeting Meta specifically, and there is no reported evidence that the stolen code was modified, published, or weaponized. No attribution, malware, CVE, or network indicators of compromise have been published. Meta's response included blocking attacker access, rotating secrets exposed in the repositories, strengthening product security, implementing controls to prevent user data being stored in GitHub, deploying proactive scans for accidentally uploaded user information, coordinating with GitHub on additional protections, notifying regulators (unnamed), and offering affected individuals credit monitoring and dark web monitoring.

Defensive relevance: this is a credential-based initial access (T1078) followed by collection from a code repository (T1213.003). Because Meta rotated secrets found in the repositories, secrets stored in code (T1552.001) are a downstream exposure risk, although use of those secrets by the attacker was not reported. Lessons are the hygiene of source-control accounts (phishing-resistant MFA, session/token controls, anomaly detection on bulk clone/download activity), keeping secrets and personal data out of repositories, and secret scanning. Coverage rests on SecurityWeek and derivative reports; Meta's own notification text was not available to this analysis, and BeaconBeagle correlation was not applicable because no network IOCs were disclosed.

MITRE ATT&CK techniques by tactic:

Initial Access

T1078 Valid Accounts

Persistence

T1078 Valid Accounts

Privilege Escalation

T1078 Valid Accounts

Defense Evasion

T1078 Valid Accounts

Collection

T1213.003 Code Repositories

Credential Access

T1552.001 Credentials In Files

Remediation for Meta GitHub breach

Immediate actions

  • Rotate all secrets, tokens and API keys stored in or ever committed to source repositories
  • Revoke and reissue developer credentials, personal access tokens and SSH keys for affected accounts
  • Review GitHub audit logs for anomalous logins, new tokens/keys and bulk clone or archive-download activity
  • Run secret scanning and PII scanning across all repositories and their history

Workarounds

  • Restrict repository access by IP allow-list or conditional access where feasible
  • Offer credit and dark web monitoring to affected individuals (as Meta did)

Longer-term hardening

  • Enforce phishing-resistant MFA (FIDO2/passkeys) and SSO with short-lived sessions for all source-control accounts
  • Prohibit personal data in repositories and enforce with pre-commit and push-protection scanning
  • Alert on unusual repository access volume, new geographies/ASNs and off-hours cloning
  • Use synthetic test data that is verifiably non-personal

Affected and fixed versions

  • Meta — Meta GitHub-hosted source code repositories (specific repositories and products not disclosed)

Timeline of Meta GitHub breach

  • Approximate window: locked-ioc-8's security team detects the intrusion 'a few weeks' before the 2026-10-13 disclosure and cuts off attacker access (exact detection and intrusion-start dates undisclosed).
  • locked-ioc-8's investigation with locked-ioc-7 finds no indication the attackers targeted locked-ioc-8 specifically; locked-ioc-8 states this was not a vulnerability in any locked-ioc-8 product and internal systems were not breached; no evidence of code modification, publication or weaponization was reported.
  • locked-ioc-8 reports it blocked attacker access, rotated exposed secrets, strengthened product security, added controls and proactive scans for user data uploaded to locked-ioc-7, is working with locked-ioc-7 on added protections, and is notifying regulators; locked-ioc-7 and law enforcement were engaged.
  • locked-ioc-8 states about 99% of the stolen data was product source code and about 1% personal information that was mistakenly placed in the repositories against policy; a significant portion was synthetic or anonymized test data.
  • locked-ioc-8 begins notifying approximately 1,500 affected individuals, mostly U.S.-based locked-ioc-8 employees, and offers credit monitoring and dark web monitoring.
  • locked-ioc-8 confirms to locked-ioc-6 that attackers used stolen employee credentials to access some of its locked-ioc-5.
  • Cyber Security News and Cyber Kendra publish follow-up coverage; no attribution, IOCs or details on how the credentials were stolen or whether MFA was enabled are provided.

Sources cited for Meta GitHub breach

Previous and next in data breach

Drafted by the Threadlinqs research pipeline from 8 sources; published ; last updated —.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

[ 0 threats ] [ 0 det ] [ CRIT: 0 ] [ HIGH: 0 ]
// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level

Latest Threats