[News](/news) LiteLLM Investigating Cyberattack [Blog](/blog)[Breaches](/breaches)[Resources](/resources)[News](/news) # LiteLLM Investigating Cyberattack ##### [UpGuard Team](/team/upguard) March 25, 2026 ### Key facts: litellm.ai data breach * **Date reported:** March 24, 2026. * **Unauthorized access identified:** March 2026 (approximate). * **Target entity:** litellm.ai (LiteLLM Python package). * **Source of breach:** TeamPCP hacking group (supply chain attack). * **Data types:** Credentials, authentication secrets, SSH keys, and cloud credentials. * **Status:** Malicious versions removed from PyPI; clean version 1.82.6 recommended. * **Severity:** High; the infostealer harvests sensitive developer and cloud infrastructure secrets. ## What happened in the litellm.ai data breach? litellm.ai (litellm.ai) was the target of a supply chain attack involving the TeamPCP hacking group. The incident, reported on March 24, 2026, centered on the compromise of the popular LiteLLM Python package hosted on the Python Package Index (PyPI). The TeamPCP hacking group compromised versions 1.82.7 and 1.82.8 to deploy an infostealer that harvests sensitive data from hundreds of thousands of devices. This malicious payload specifically targets credentials and authentication secrets, including SSH keys and cloud credentials. The severity is high because these stolen secrets can grant attackers persistent access to cloud environments and developer infrastructure. While the malicious versions have been removed and replaced by the clean version 1.82.6, affected users face significant risks of unauthorized access and data theft. ## Who is behind the incident? The TeamPCP hacking group is responsible for the compromise of the LiteLLM package on PyPI. While specific details about the group's origin or geographic location are not currently disclosed, their tactics involve supply chain attacks aimed at distributing [infostealer malware](https://www.upguard.com/blog/why-infostealer-malware-demands-a-new-defense-strategy). By targeting a widely used library like LiteLLM, the group demonstrated a focus on harvesting high-value developer credentials and cloud access keys. Their activities indicate a sophisticated approach to bypassing traditional security measures by injecting malicious code directly into the software development lifecycle. TeamPCP's campaign highlights the growing threat of package repository compromises. ## Impact and risks for litellm.ai customers For developers and organizations using LiteLLM, the primary risk involves the theft of sensitive authentication secrets. The infostealer's ability to harvest SSH keys and cloud credentials could lead to unauthorized access to internal systems, identity theft, and potential service disruptions. Because these credentials often provide broad access to cloud infrastructure, the breach could result in significant lateral movement by attackers within compromised environments. Supply chain attacks typically lead to widespread credential exposure and long-term security vulnerabilities. Affected users should immediately rotate all secrets, revoke potentially compromised keys, and monitor account logs for suspicious activity. Maintaining transparency about the incident and following rigorous remediation protocols are essential for minimizing the long-term impact on organizational security. ## How to protect against similar security incidents Given the nature of this supply chain attack on litellm.ai, users must act quickly to purge malicious package versions and secure any credentials that may have been harvested by the infostealer. * ‍**Revert to clean package versions.** Immediately uninstall LiteLLM versions 1.82.7 and 1.82.8. Downgrade to the verified clean version 1.82.6 or later as specified by the maintainers.**‍** * **Rotate credentials and secrets.** Revoke and regenerate all SSH keys, cloud access tokens, and API keys present on affected systems. Assume any secret stored in environment variables or configuration files on compromised devices is at risk.**‍** * **Implement software composition analysis.** Use SCA tools to monitor project dependencies for known malicious versions or vulnerabilities. Establish a process for verifying the integrity of third-party packages before deployment in production environments.**‍** * **Monitor for unauthorized access.** Audit cloud provider logs and SSH access logs for any unusual activity originating from affected developer machines. Deploy continuous attack surface management to identify exposed credentials or unauthorized infrastructure changes. *Prompt action in rotating secrets and updating dependencies is the most effective way to mitigate the risks associated with this supply chain incident.* ## Frequently asked questions ### What happened in the litellm.ai security breach? TeamPCP claimed responsibility for a security attack on litellm.ai (litellm.ai) in March 2026. The incident was first reported on March 24, 2026. ### When did the litellm.ai breach occur? The litellm.ai breach was publicly reported on March 24, 2026. TeamPCP referenced the incident around that time, but the attack may have occurred earlier. ### What data was exposed? The litellm.ai incident involved the exposure of credentials, authentication secrets, SSH keys, and cloud credentials. TeamPCP deployed an infostealer specifically to harvest these categories of data from affected devices. ### Is my personal information at risk? If you interacted with litellm.ai, there's a possibility your personal information could be affected. Similar incidents often involve email addresses, login details, or financial records. Stay alert for updates and take precautionary measures to secure your accounts. ### How can I protect myself after a data breach? • Uninstall LiteLLM versions 1.82.7 and 1.82.8 immediately.\ • Rotate all SSH keys and cloud credentials.\ • Enable phishing-resistant MFA on all sensitive accounts.\ • Monitor financial and cloud service accounts for unauthorized activity.\ • Use breach monitoring tools to track credential exposure. ### What steps should companies take after being breached? The maintainers of LiteLLM have removed the malicious versions from PyPI and are investigating the incident. Organizations should secure systems, notify affected parties, provide guidance on protective actions, review security measures, and deploy attack surface management. *This cybersecurity news article is powered by *[*UpGuard Breach Risk*](https://www.upguard.com/product/breach-risk)* — continuous attack surface monitoring for your organisation and supply chain.* ## How secure is LiteLLM? LiteLLM is a San Francisco-based AI infrastructure company that provides an open-source gateway and Python SDK designed to unify over 100 large language model APIs into a single interface, allowing developers to simplify model integration, cost tracking, and failover management. * View our free preliminary report on LiteLLM’s security posture * 13 risk factors, including email security, SSL, DNS health, open ports and common vulnerabilities [View LiteLLM's score](/security-report/litellm) [View score](/security-report/litellm) [https://www.litellm.ai/](/security-report/litellm) ### Sign up for our newsletter UpGuard's monthly newsletter cuts through the noise and brings you what matters most: our breaking research, in-depth analysis of emerging threats, and actionable strategic insights. ## Latest news Stay up-to-date with the latest news in cybersecurity. [](/news/hugging-face-data-breach-2026-07-20) #### [Hugging Face data breach: key facts and what we know so far](/news/hugging-face-data-breach-2026-07-20) A data breach involving Hugging Face was reported in July 2026. See incident details, impact on customers, and recommended security measures. ##### [UpGuard Team](/team/upguard) July 20, 2026 [](/news/ey-data-breach-2026-07-19) #### [EY data breach: what happened and what's at risk](/news/ey-data-breach-2026-07-19) A data breach involving EY was reported in July 2026. See incident details, impact on customers, and recommended security measures. ##### [UpGuard Team](/team/upguard) July 15, 2026 [](/news/south-florida-injury-convenient-care-data-breach-2026-07-16) #### [South Florida Injury & Convenient Care data breach exposes names and Social Security numbers](/news/south-florida-injury-convenient-care-data-breach-2026-07-16) A data breach involving South Florida Injury & Convenient Care was reported in July 2026. See incident details, impact on customers, and security measures. ##### [UpGuard Team](/team/upguard) July 15, 2026 [](/news/park-west-psychology-data-breach-2026-07-16) #### [Park West Psychology data breach: what happened and what's at risk](/news/park-west-psychology-data-breach-2026-07-16) A data breach involving Park West Psychology was reported in July 2026. See incident details, impact on customers, and recommended security measures. ##### [UpGuard Team](/team/upguard) July 15, 2026 [](/news/fall-river-municipal-credit-union-data-breach-2026-07-16) #### [Fall River Municipal Credit Union data breach: what happened and what's at risk](/news/fall-river-municipal-credit-union-data-breach-2026-07-16) A data breach involving Fall River Municipal Credit Union was reported in July 2026. See incident details, impact on customers, and recommended security measures. ##### [](#) July 14, 2026 [](/news/npcil-data-breach-2026-07-16) #### [NPCIL data breach: World Leaks claims exposure of Kudankulam Nuclear Power Plant files](/news/npcil-data-breach-2026-07-16) A data breach involving NPCIL was reported in July 2026. See incident details, impact on customers, and recommended security measures. ##### [UpGuard Team](/team/upguard) July 14, 2026 [View all news](/news) ## Protect your organization Get in touch or book a free demo. [Contact sales](/demo) [Free demo](/demo) ##### Free instant security score ## How secure is your organization? Request a free cybersecurity report to discover key risks on your website, email, network, and brand. * Instant insights you can act on immediately * Hundreds of risk factors including email security, SSL, DNS health, open ports and common vulnerabilities [Free score](/instant-security-score)