INSPECTING ARCHIVED INTELLIGENCE (OUTDATED VERSION).
LiteLLM Master Key Leaked Allowing Attackers to Hijack GitHub Releases
|
2026-10-24 11:03
HIGH
LOW
DATA BREACH
CRITICAL INFRASTRUCTURE & OT
Executive Summary
AI-generated
An unknown attacker gained access to the Agent Platform master key, which was accidentally published to a public GitHub repository in late May 2026. The attacker used this unrevoked key on October 10, 2026, to gain owner-level access to all repositories across an organization and perform additional credential exfiltration. The attacker extracted a GitHub CLI OAuth token associated with the organization's account, which allowed them to possess owner-level access to all repositories. This attack worked by exploiting the relaxed release approval process that occurred earlier in May 2026, allowing any two engineers to trigger and approve releases without authorization. As of October 21, 2026, the affected organization had reported a supply chain attack, with PyPI quarantining LiteLLM version 1.107.0 and Docker Hub removing or repointing tags related to this release.
Technical Mitigations AI-generated
• Patch the LiteLLM Agent Platform to revoke the exposed master key and prevent future unauthorized access.
• Monitor GitHub repositories for suspicious activity, such as commits from throwaway branches or unauthorized webhook endpoints.
• Implement strict branch protection and release pipeline controls to prevent unauthorized modifications.
Technical Observables Login Required
Indicators of Compromise (IoCs) Protected
Raw threat telemetry, malicious IP addresses, file hashes, and direct VirusTotal correlation are restricted to authenticated users.
li•••••.dev
69.48.•••.•••
Intelligence Metadata
Actors / Malware / CVEs / Campaigns
Target & Sectors
NORTH_AMERICA
NORTH_AMERICA
technologytechnology
Incident Timeline
October 21, 2025
Attackers exploited a vulnerability in the LiteLLM model to obtain its master key, which they then used to steal GitHub tokens and hijack release pipelines.
Click on any entity below to view its context and source!
industry
Technology
General Document Context
late May 2026
LiteLLM on Friday revealed that an Agent Platform master key accidentally published to a public GitHub repository in late May 2026 ultimately allowed the threat actor to
push a malicious version
of the popular Python package.
Click on any entity below to view its context and source!
tactic
T1059.006 - Python
LiteLLM on Friday revealed that an Agent Platform master key accidentally published to a public GitHub repository in late May 2026 ultimately allowed the threat actor to
push a malicious version
of the popular Python package.
May 2026
The leaked LiteLLM master key in May 2026 was used to hijack a release pipeline, allowing attackers to steal GitHub tokens.
Click on any entity below to view its context and source!
organisation
Leaked
Leaked master key in May 2026
"This restriction was relaxed earlier this week to accommodate the temporary push for more frequent releases and bugfix backports, allowing any two engineers to trigger and approve a release.
organisation
API
LiteLLM also said the workflows are designed to identify and exfiltrate additional keys and tokens, specifically focusing on upstream inference API tokens.
organisation
the Workflows Emerge
More Details About the Workflows Emerge
Socket, which
published
its own investigation into the malicious workflows, said it recovered two malicious commits that each added a workflow at ".github/workflows/cache.yml."
May 27, 2026
A configuration file was committed to the public LiteLLM-Labs/lite-harness repository on May 27, 2026, inadvertently containing the master key to a deployment of the LiteLLM Agent Platform.
Click on any entity below to view its context and source!
industry
Technology
General Document Context
October 10, 2026
An attacker used a leaked LiteLLM master key to gain unauthorized access, steal GitHub tokens and hijack the release pipeline.
Click on any entity below to view its context and source!
industry
Technology
General Document Context
observable
69.48.229.143
"
LiteLLM said the unknown attacker (operating from "69.48.229[.]143") used this unrevoked key on October 10, 2026, to gain access to the Agent Platform and extracted a GitHub CLI OAuth token the service was configured with to allow self-hosted cloud agents to file pull requests to GitHub repositories.
organisation
CLI
"
LiteLLM said the unknown attacker (operating from "69.48.229[.]143") used this unrevoked key on October 10, 2026, to gain access to the Agent Platform and extracted a GitHub CLI OAuth token the service was configured with to allow self-hosted cloud agents to file pull requests to GitHub repositories.
organisation
UTC
The sequence of actions performed by the threat actor following this step is as follows -
October 10, 2026, between 10:11 and 10:12 p.m. UTC:
infrastructure
1.107.0
What the Malicious Version of the Package Does
Multiple reports have
revealed
that the files bundled with LiteLLM version 1.107.0 are engineered to execute automatically when Python starts, even if an application did not import LiteLLM.
organisation
GitHub
The token, per the developers of the open-source large language model (LLM) gateway, was "not scoped down" and associated with its GitHub organization owner's account.
organisation
CLI Oauth
Author commits into throwaway branches on internal repositories to run workflows (BerriAI/litellm and BerriAI/litellm-auto) designed to acquire GitHub CLI Oauth tokens associated with two engineers with maintainer privileges via an attacker-controlled webhook[.]site endpoint.
organisation
The Hacker News
"Neither commit was pushed to a branch that still exists," Socket security researcher Karlo Zanki said in a report shared with The Hacker News.
organisation
Linear
Both workflows, as highlighted by LiteLLM, are equipped to harvest GitHub tokens and API keys for the organization's Linear workspace, LiteLLM deployment, and Together AI, along with the identifier and signing key of a GitHub App.
organisation
SSH
Compromised LiteLLM 1.107.0 execution flow
The stealer also collects system information, SSH files, Git and GitLab credentials and tokens, environment variables, and tokens associated with npm, PyPI, pip, Yarn, Maven, Gradle, Composer, RubyGems, and Cargo.
organisation
GitLab
Compromised LiteLLM 1.107.0 execution flow
The stealer also collects system information, SSH files, Git and GitLab credentials and tokens, environment variables, and tokens associated with npm, PyPI, pip, Yarn, Maven, Gradle, Composer, RubyGems, and Cargo.
organisation
npm
Compromised LiteLLM 1.107.0 execution flow
The stealer also collects system information, SSH files, Git and GitLab credentials and tokens, environment variables, and tokens associated with npm, PyPI, pip, Yarn, Maven, Gradle, Composer, RubyGems, and Cargo.
organisation
RubyGems
Compromised LiteLLM 1.107.0 execution flow
The stealer also collects system information, SSH files, Git and GitLab credentials and tokens, environment variables, and tokens associated with npm, PyPI, pip, Yarn, Maven, Gradle, Composer, RubyGems, and Cargo.
organisation
Cargo
Compromised LiteLLM 1.107.0 execution flow
The stealer also collects system information, SSH files, Git and GitLab credentials and tokens, environment variables, and tokens associated with npm, PyPI, pip, Yarn, Maven, Gradle, Composer, RubyGems, and Cargo.
October 12
Attackers exploited the extended collection of secrets on GitHub to steal tokens and hijack release pipelines.
Click on any entity below to view its context and source!
organisation
AWS Bedrock
The October 12 run extends the collection to more secrets, including API keys for Anthropic, OpenAI, AWS Bedrock, Tavily, and GitGuardian's secret-scanning tenant.
organisation
GitGuardian
The October 12 run extends the collection to more secrets, including API keys for Anthropic, OpenAI, AWS Bedrock, Tavily, and GitGuardian's secret-scanning tenant.
October 13, 2026
Attackers exploited a vulnerability in LiteLLM deployments to activate malicious code, potentially exposing GitHub tokens and hijacking release pipelines.
Click on any entity below to view its context and source!
industry
Technology
General Document Context
organisation
Microsoft Security Research
"Together, these mechanisms extended potential exposure beyond active LiteLLM deployments: simply starting Python in an affected environment could activate the malicious code, even when LiteLLM itself was never imported," the Microsoft Security Research team
said
.
October 21, 2026
Attackers exploited the leaked LiteLLM master key to hijack a GitHub token, which was then used to target and hijack a release pipeline.
Click on any entity below to view its context and source!
industry
Technology
General Document Context
infrastructure
1.107.0
UTC: LiteLLM version 1.107.0 gets published to PyPI on
October 21, 2026, 11:46 a.m.
UTC: PyPI quarantines LiteLLM version 1.107.0.
organisation
Upwind
UTC: Supply chain attack reported by Upwind researcher Lidor Machluf.
October 21, 2026, 1:23 p.m.
organisation
PyPI
UTC: PyPI quarantines LiteLLM version 1.107.0.
2026/10/24
Attackers exploited the leaked LiteLLM master key to gain unauthorized access, steal GitHub tokens, and subsequently hijack a release pipeline on October 24, 2026.
Click on any entity below to view its context and source!
organisation
Hijack Release Pipeline
Leaked LiteLLM Master Key Let Attackers Steal GitHub Tokens and Hijack Release Pipeline.
Tactical Metrics
Metrics
infrastructure
1.107.0
Software Version
Click for context!
UTC: LiteLLM version 1.107.0 gets published to PyPI on
October 21, 2026, 11:46 a.m.
UTC: PyPI quarantines LiteLLM version 1.107.0.
What the Malicious Version of the Package Does
Multiple reports have
revealed
that the files bundled with LiteLLM version 1.107.0 are engineered to execute automatically when Python starts, even if an application did not import LiteLLM.
Intelligence Sources
The Hacker News
2026-10-24
Unpublish from Social Media?
Are you sure you want to delete this podcast video from all synchronized social networks (YouTube, Facebook, Threads)?
Important:
Due to Meta API restrictions, Instagram Reels cannot be deleted automatically via API by third-party apps.
View Profile to Delete Manually
View Profile to Delete Manually
Tactical Intelligence
Report Intelligence Issue
Podcast Options
Generate
Incident Version History
CURRENT VERSION
Last Updated: 2026-10-25T07:08
Comprehensive Tactical Telemetry
Highly Correlated Entities
20x
organisation
Identified Entity
CLI
entity
10x
timeline
Temporal Reference
May 27, 2026
date
Contextual Telemetry
Context Block
5 METRICS
target region
Target Country
United States
country
industry
Targeted Sector
Technology
sector
tactic
Cyber Operation Type
Exfiltration
tactic
infrastructure
Software Version
1.107.0
version
tactic
MITRE ATT&CK Technique
T1059.006 - Python
technique
Click on any entity below to view its context in the main text!
Selective Unpublish
Select the networks you want to remove this post from. The system will try to delete the real post through the API and clean the database so you can publish it again.
By navigating this website, you accept the use of strictly necessary technical cookies for session security and basic platform functionality. We do not use tracking or advertising cookies. Read our Privacy Policy.