CISA list

● PyPI package

litellm

Litellm has 44 security advisories. Four CVEs named in them are on CISA's list of exploited vulnerabilities. The newest release any of them names as the fix is 1.104.0rc2.

As of , litellm has 44 security advisories. Four CVEs named in them are on CISA's list of exploited vulnerabilities. The newest release any of them names as the fix is 1.104.0rc2.

Your version, and the version to upgrade to

Find the row your installed version falls in. The second column is the first release after it that none of the advisories lists as affected; a release outside every row is one no advisory lists. That is what the advisories say, not a promise that a release is free of flaws.
If you runUpgrade toAdvisories
From 1.104.0rc1 before 1.104.0rc21.104.0rc21 advisory
From 1.103.0 before 1.103.11.103.11 advisory
From 1.102.0 before 1.102.21.102.21 advisory
From 1.101.0 before 1.101.31.101.31 advisory
Before 1.100.41.100.444 advisoriesOne is on CISA's list

Security advisories, newest first

Security advisories, newest first
AdvisoryWhat it isAffectsFixed inPublished
GHSA-g5ff-637f-6q2mLiteLLM: internal_user_viewer arbitrary local file read via `vertex_ai_credentials`High severityBefore 1.95.01.95.0
GHSA-hhww-mrg2-969hLiteLLM vulnerable to reflected XSS in /sso/debug/callbackModerate severityFrom 1.65.5 before 1.85.01.85.0
CVE-2026-114612GHSA-7hp6-4w63-5g45LiteLLM: Privilege Escalation to Proxy Admin via Cross-Domain Reuse of the Salt KeyHigh severityFrom 1.104.0rc1 before 1.104.0rc2; From 1.103.0 before 1.103.1; From 1.102.0 before 1.102.2; From 1.101.0 before 1.101.3; From 1.91.0 before 1.100.41.100.4, 1.101.3, 1.102.2, 1.103.1, 1.104.0rc2
CVE-2026-84377GHSA-3cv6-jpf6-8222, PYSEC-2026-4066LiteLLM: Authenticated SSRF and provider-credential exfiltration via unvalidated request-body routing parametersModerate severityFrom 1.96.0 before 1.96.2; From 1.95.0 before 1.95.1; From 1.94.0 before 1.94.3; From 1.93.0 before 1.93.2; From 1.92.0 before 1.92.2; From 1.91.0 before 1.91.5; From 1.90.0 before 1.90.7; From 1.89.0 before 1.89.7; Before 1.88.61.88.6, 1.89.7, 1.90.7, 1.91.5, 1.92.2, 1.93.2 and 3 more
CVE-2026-59823GHSA-hx8v-g79f-8w5f, PYSEC-2026-4070LiteLLM Proxy has server-side request forgery via the `user_config` request parameterModerate severityBefore 1.83.91.83.9
CVE-2026-37004GHSA-6wvf-77m9-58rm, PYSEC-2026-3861LiteLLM vulnerable to server-side template injection in the /prompts/test endpointCritical severityBefore 1.83.71.83.7
CVE-2026-59821GHSA-72m8-9m7m-h278, PYSEC-2026-3478LiteLLM: Custom Code Guardrails production endpoints bypass code safety checksLow severityBefore 1.82.01.82.0
CVE-2026-59822GHSA-7488-6r32-c95q, PYSEC-2026-3479On CISA's list: CVE-2026-59822LiteLLM: MCP Authentication Bypass via OAuth2 Passthrough FallbackHigh severityBefore 1.84.01.84.0
CVE-2026-59819GHSA-4g5m-c9r5-49xf, PYSEC-2026-3476LiteLLM: Local file read via request-supplied OIDC file referencesLow severityBefore 1.83.101.83.10
CVE-2026-59820GHSA-5jmr-gcrj-2c9q, PYSEC-2026-3477LiteLLM: Arbitrary file write via path traversal in Skills archive extractionModerate severityBefore 1.83.71.83.7
CVE-2026-12798GHSA-c693-x898-5g4h, PYSEC-2026-4069BerriAI litellm has SSRF via Unvalidated spec_path URL in MCP OpenAPI Spec LoaderLow severityUp to and including 1.82.2No fixed version listed
CVE-2026-12799GHSA-m2v5-74w2-qhcj, PYSEC-2026-4072BerriAI litellm: UI User Enumeration leads to System-Wide Information DisclosureLow severityUp to and including 1.82.2No fixed version listed
CVE-2026-12797GHSA-p897-vf7j-f5h8, PYSEC-2026-4074BerriAI litellm has Security Feature Bypass in BannedKeywords and AzureContentSafety Guardrails via call_type Mismatch on Async EndpointsLow severityUp to and including 1.82.5No fixed version listed
CVE-2026-12796GHSA-w2mh-qq9q-453x, PYSEC-2026-4076BerriAI litellm: SSO Login Does Not Invalidate Previous UI Session TokensLow severityUp to and including 1.82.2No fixed version listed
CVE-2026-12795GHSA-j37q-q7p9-vpwm, PYSEC-2026-4071LiteLLM: SSO Debug Flow Has Improper AuthenticationModerate severityUp to and including 1.82.2No fixed version listed
CVE-2026-12773GHSA-4jcj-7x88-m979, PYSEC-2026-4067LiteLLM: MCP Proxy Has Improper AuthenticationModerate severityBefore 1.84.01.84.0
CVE-2026-12770GHSA-6qr3-3g89-m4jj, PYSEC-2026-4068LiteLLM: Admin Key Handler Has Improper AuthorizationLow severityUp to and including 1.63.1No fixed version listed
CVE-2026-12772GHSA-mf52-j94g-746m, PYSEC-2026-4073LiteLLM: PROXY_ADMIN database API Key Generator Has Insufficient Session ExpirationLow severityUp to and including 1.82.2No fixed version listed
CVE-2026-12771GHSA-qmf3-4767-5fg3, PYSEC-2026-4075LiteLLM: M2M JWT Handler Has Improper AuthorizationLow severityUp to and including 1.82.2No fixed version listed
CVE-2026-49468GHSA-4xpc-pv4p-pm3w, PYSEC-2026-388LiteLLM: Authentication Bypass via Host Header InjectionCritical severityBefore 1.84.01.84.0
CVE-2026-47102GHSA-wpfp-gwwc-vwq6, PYSEC-2026-2600LiteLLM allows a user to modify their own user_role via the /user/update endpointHigh severityBefore 1.83.101.83.10
CVE-2026-47101GHSA-qrc4-49gv-mv9m, PYSEC-2026-2598LiteLLM allows an authenticated internal_user to create API keys with access to routes that their role does not permitHigh severityBefore 1.83.141.83.14
CVE-2026-40217GHSA-wxxx-gvqv-xp7p, PYSEC-2026-2601On CISA's list: CVE-2026-40217LiteLLM has a sandbox escape in custom-code guardrailHigh severityFrom 1.81.8 before 1.83.101.83.10
CVE-2026-42271GHSA-v4p8-mg3p-g94g, PYSEC-2026-2599On CISA's list: CVE-2026-42271LiteLLM: Authenticated command execution via MCP stdio test endpointsHigh severityFrom 1.74.2 before 1.83.71.83.7
CVE-2026-42208GHSA-r75f-5x8p-qvmc, PYSEC-2026-391On CISA's list: CVE-2026-42208LiteLLM has SQL Injection in Proxy API key verificationCritical severityFrom 1.81.16 before 1.83.71.83.7
CVE-2026-42203GHSA-xqmj-j6mv-4862, PYSEC-2026-2602LiteLLM: Server-Side Template Injection in /prompts/test endpointHigh severityFrom 1.80.5 before 1.83.71.83.7
GHSA-69x8-hrgq-fjj8LiteLLM: Password hash exposure and pass-the-hash authentication bypassHigh severityBefore 1.83.01.83.0
CVE-2026-35030GHSA-jjhc-v7c2-5hh6, PYSEC-2026-390LiteLLM: Authentication bypass via OIDC userinfo cache key collisionCritical severityBefore 1.83.01.83.0
CVE-2026-35029GHSA-53mr-6c8q-9789, PYSEC-2026-2597LiteLLM: Privilege escalation via unrestricted proxy configuration endpointHigh severityBefore 1.83.01.83.0
PYSEC-2026-2Two litellm versions published containing credential harvesting malwareFrom 1.82.7 up to and including 1.82.8No fixed version listed
CVE-2025-0330GHSA-879v-fggm-vxw2, PYSEC-2026-1543LiteLLM Has a Leakage of Langfuse API KeysHigh severityUp to and including 1.52.1No fixed version listed
CVE-2025-0628GHSA-fjcf-3j3r-78rp, PYSEC-2026-1546LiteLLM Has an Improper Authorization VulnerabilityHigh severityBefore 1.61.151.61.15
CVE-2024-9606GHSA-g5pg-73fc-hjwq, PYSEC-2026-1548LiteLLM Reveals Portion of API Key via a Logging FileHigh severityBefore 1.44.121.44.12
CVE-2024-8984GHSA-fh2c-86xm-pm2x, PYSEC-2026-1545LiteLLM Vulnerable to Denial of Service (DoS) via Crafted HTTP RequestHigh severityBefore 1.56.21.56.2
CVE-2024-6825GHSA-53gh-p8jc-7rg8, PYSEC-2026-1541LiteLLM Vulnerable to Remote Code Execution (RCE)High severityFrom 1.40.3.dev2 up to and including 1.40.12No fixed version listed
CVE-2024-10188GHSA-gw2q-qw9j-rgv7, PYSEC-2026-1549LiteLLM Vulnerable to Denial of Service (DoS)High severityBefore 1.53.1.dev11.53.1.dev1
CVE-2024-6587GHSA-g26j-5385-hhw3, PYSEC-2026-1547LiteLLM Server-Side Request Forgery (SSRF) vulnerabilityHigh severityBefore 1.44.81.44.8
CVE-2024-5751GHSA-gppg-gqw8-wh9g, PYSEC-2026-389litellm vulnerable to remote code execution based on using eval unsafelyCritical severityBefore 1.40.161.40.16
CVE-2024-5710GHSA-qqcv-vg9f-5rr3, PYSEC-2026-1551litellm vulnerable to improper access control in team managementModerate severityBefore 1.40.151.40.15
CVE-2024-4888GHSA-3xr8-qfvj-9p9j, PYSEC-2026-1540Arbitrary file deletion in litellmHigh severityBefore 1.35.361.35.36
CVE-2024-4890GHSA-8j42-pcfm-3467, PYSEC-2026-1544SQL injection in litellmModerate severityUp to and including 1.27.14No fixed version listed
CVE-2024-5225GHSA-h6m6-jj8v-94jj, PYSEC-2026-1550SQL injection in litellmModerate severityBefore 1.40.01.40.0
CVE-2024-4264GHSA-7ggm-4rjg-594w, PYSEC-2026-1542litellm passes untrusted data to `eval` function without sanitizationHigh severityUp to and including 1.28.11No fixed version listed
CVE-2024-2952GHSA-46cm-pfwv-cgf8, PYSEC-2026-387LiteLLM has Server-Side Template Injection vulnerability in /completions endpointCritical severityBefore 1.34.421.34.42

Other PyPI packages

Sources
  • Package advisories from the GitHub Advisory Database (CC BY 4.0). Version ranges are reworded as sentences (source). Licence: CC BY 4.0
  • Package advisories from the PyPA advisory database (CC BY 4.0). Version ranges are reworded as sentences (source). Licence: CC BY 4.0