LesenForschungRadarAnlageframework
Anmelden / Registrieren
← Zurück zur Lektüre
NewsThema · LiteLLM供应链安全Wesentlich2 Min. LesezeitDisconfirmAI Redaktion

AI Proxy Library LiteLLM Hit by Supply Chain Attack; Malicious v1.107.0 Steals Cloud and Model API Keys

Popular AI proxy library LiteLLM suffered a supply chain attack with malicious version v1.107.0 implanting credential stealers, threatening cloud and LLM API keys.

2026-10-21 21:44 GMT+8

Security firm Wiz Research disclosed on October 21, 2026, that the popular Python LLM proxy library LiteLLM was targeted by a supply chain attack. Attackers published a malicious version, v1.107.0, which remained available on PyPI for approximately two hours before being quarantined. Simultaneously, related image tags on GHCR and Docker Hub (including latest, main-latest, etc.) were trojanized until they were removed around 13:53 UTC. The most recent safe version is v1.106.2.

According to analyses by Wiz and Datadog Security Labs, the malicious package extracts three files (cook.pth, cook.py, and sitecustomize.py) into the site-packages root directory upon installation. These files are designed to execute automatically whenever a Python interpreter starts, triggering even if LiteLLM is not imported. Specifically, cook.py contains a credential stealer named privcollect, which harvests sensitive data across 16 modules targeting browser credentials from Chrome/Firefox, AWS/GCP/Azure cloud credentials, SSH/GPG keys, and API keys for LLM providers like OpenAI and Anthropic stored in environment variables or config files. It also directly queries the LiteLLM Proxy's PostgreSQL database to extract upstream provider credentials.

The harvested data is encrypted with AES-256-CBC and exfiltrated via HTTP POST requests to the attacker-controlled lookalike domain www.litellm[.]dev/healthcheck ↗. The sitecustomize.py file includes an hourly backdoor mechanism checking a GitHub dead drop to download second-stage RAT payloads, although no active payload was observed in the dead drop at the time of analysis. Wiz noted that this dual-technique approach using both .pth files and sitecustomize.py appears to be the first documented case in the wild.

Anmelden / Registrieren
LesenForschungRadarAnlageframework

LiteLLM developer BerriAI stated that attackers used stolen GitHub CLI OAuth tokens from three engineers' accounts to trigger the legitimate build pipeline against a malicious fork repository (Berri-AI/litellm, noting the hyphen difference). Datadog added that this marks the second supply chain compromise of LiteLLM in 2026, following earlier incidents affecting versions 1.82.7 and 1.82.8 linked to TeamPCP activity, though no connection between the two compromises has been established yet.

Affected users should immediately manually remove the cook.pth, cook.py, and sitecustomize.py files from the site-packages directory, revert to v1.106.2, and rotate all potentially exposed cloud, AI, and browser credentials from a clean machine.

Quellen

  • wiz.io↗
  • securitylabs.datadoghq.com↗
Thema

LiteLLM供应链安全

【背景】LiteLLM是广泛使用的开源AI模型网关,用于统一调用多家大模型API并管理密钥。2026年,该项目连续遭遇两次重大供应链攻击,暴露出开源基础设施在CI/CD流程与凭证管理上的脆弱性。

5 Schritte2 BeiträgeAktualisiert 2026-10-26 05:43

Wie es dazu kam

  1. 2026-03-24LiteLLM v1.82.7和v1.82.8因Trivy依赖被劫持遭供应链投毒,恶意代码窃取SSH密钥与云凭证,随后被下架。docs.litellm.ai ↗
  2. 2026-05-27LiteLLM团队在公开仓库误提交包含Agent Platform主密钥的配置文件,虽快速撤回但未及时轮换该密钥,埋下隐患。docs.litellm.ai ↗
  3. 2026-10-10攻击者利用未轮换的GitHub CLI OAuth令牌访问LiteLLM Agent Platform,进而获取组织所有者权限及多名工程师的维护者令牌。docs.litellm.ai ↗

Was noch ungeklärt ist

  • 攻击者最初如何具体利用Claude Code生成的日志泄露来获取内部部署访问权限?
  • 10月攻击是否与3月的Trivy供应链攻击由同一威胁行为者发起?
  • LiteLLM新实施的Cosign签名验证机制能否有效防止未来的静默篡改?

Auch in diesem Thread

  • Open-source AI model gateway LiteLLM hit by supply chain attack; official release pipeline compromised to steal keys2026-10-22
Vollständiges Thema lesen →
Recherche fortsetzen →