high CVSS 7.2

CVE-2026-59821·Python vulnerability

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.82.0-stable, LiteLLM's Custom Code Guardrails production create and update paths did not apply the same sandboxing and validation used by the test endpoint, allowing a privileged user with access to create or update guardrails to submit custom Python code that executed in the LiteLLM proxy environment and could expose secrets available to the process. This issue is fixed in version 1.82.0-stable.

Severity
high
Software
Python
Fixed in
1.82.0
Published
2026-07-08

Affected versions

Until: 1.82.0

Fixed in: 1.82.0

How to fix this CVE

Upgrade LiteLLM to version 1.82.0-stable or later to apply mandatory sandboxing and validation to custom code guardrails in production environments. This patch ensures that guardrail code execution is restricted to the same secure sandbox used during testing, preventing privileged users from injecting malicious Python code that could access sensitive environment variables and secrets within the proxy process.

sudo dnf update python3-litellm

Defensia detects this vulnerability

WAF mitigation (if patching is not yet possible)

Add this rule to your WAF to block exploitation attempts while you schedule the patch. 

Implement API-level access controls to restrict guardrail creation and modification endpoints to a minimal set of administrative users. Deploy input validation on guardrail payloads to reject requests containing Python code keywords (exec, eval, import, __import__) at the application firewall layer before reaching the LiteLLM service.

How to check if you are affected

  1. Check the installed LiteLLM version: python3 -c "import litellm; print(litellm.__version__)"
  2. Verify if LiteLLM is running as a proxy service: ps aux | grep litellm
  3. Review guardrail configuration files for custom Python code entries: grep -r "custom_code" /etc/litellm/ 2>/dev/null || find ~ -name "*guardrail*" -type f
  4. Confirm the patch was applied by checking the version again and comparing against 1.82.0-stable: python3 -c "import litellm; print('Vulnerable' if litellm.__version__ < '1.82.0' else 'Patched')"

FAQ

What is CVE-2026-59821?

CVE-2026-59821 is a privilege escalation vulnerability in LiteLLM versions before 1.82.0 where custom Python code submitted to production guardrails bypasses sandbox restrictions, allowing high-privileged users to execute arbitrary code and exfiltrate environment secrets stored in the proxy process.

Is CVE-2026-59821 being actively exploited?

No, there is no evidence of active exploitation in the wild or public exploit code available. However, the vulnerability requires admin/privileged access to LiteLLM guardrail management, limiting the attack surface to trusted internal users.

What versions of LiteLLM are affected by CVE-2026-59821?

All versions of LiteLLM prior to version 1.82.0-stable are affected by this vulnerability.

How do I check if my server is vulnerable to CVE-2026-59821?

Run 'python3 -c "import litellm; print(litellm.__version__)"' and compare the version number to 1.82.0-stable; if your version is lower, your installation is vulnerable.

Does Defensia detect CVE-2026-59821?

Yes — Defensia's CVE advisory scanner compares installed package versions against the NVD database. If LiteLLM is installed on a monitored server, CVE-2026-59821 will appear in your dashboard with remediation steps.

Related Python CVEs

CVE-2026-34938CVSS 10PraisonAI is a multi-agent teams system. Prior to version 1.5.90, execute_code() in praisonai-agents runs attacker-controlled Python inside a three-layer sandbox that can be fully bypassed by passing a str subclass with an overridden startswith() method to the _safe_getattr wrapper, achieving arbitrary OS command execution on the host. This issue has been patched in version 1.5.90.
CVE-2026-33054CVSS 10Mesop is a Python-based UI framework that allows users to build web applications. Versions 1.2.2 and below contain a Path Traversal vulnerability that allows any user supplying an untrusted state_token through the UI stream payload to arbitrarily target files on the disk under the standard file-based runtime backend. This can result in application denial of service (via crash loops when reading non-msgpack target files as configurations), or arbitrary file manipulation. This vulnerability heavily exposes systems hosted utilizing FileStateSessionBackend. Unauthorized malicious actors could interact with arbitrary payloads overwriting or explicitly removing underlying service resources natively outside the application bounds. This issue has been fixed in version 1.2.3.
CVE-2026-28505CVSS 10Tautulli is a Python based monitoring and tracking tool for Plex Media Server. Prior to version 2.17.0, the str_eval() function in notification_handler.py implements a sandboxed eval() for notification text templates. The sandbox attempts to restrict callable names by inspecting code.co_names of the compiled code object. However, co_names only contains names from the outer code object. When a lambda expression is used, it creates a nested code object whose attribute accesses are stored in code.co_consts, NOT in code.co_names. The sandbox never inspects nested code objects. This issue has been patched in version 2.17.0.
CVE-2026-25632CVSS 10EPyT-Flow is a Python package designed for the easy generation of hydraulic and water quality scenario data of water distribution networks. Prior to 0.16.1, EPyT-Flow’s REST API parses attacker-controlled JSON request bodies using a custom deserializer (my_load_from_json) that supports a type field. When type is present, the deserializer dynamically imports an attacker-specified module/class and instantiates it with attacker-supplied arguments. This allows invoking dangerous classes such as subprocess.Popen, which can lead to OS command execution during JSON parsing. This also affects the loading of JSON files. This vulnerability is fixed in 0.16.1.
CVE-2026-8476CVSS 9.9IBM Langflow OSS 1.0.0 through 1.10.0 contain a critical remote code execution vulnerability in the disk-based caching mechanism. The AsyncDiskCache class uses Python's unsafe pickle.loads() function to deserialize cached objects from disk without validation, integrity verification, or authentication, enabling arbitrary code execution when malicious pickle payloads are processed. Attackers who can influence cached data through file system access, malicious workflow inputs, custom components, or API manipulation can achieve complete system compromise with the privileges of the Langflow server process.

References

Track CVEs across your fleet automatically

Defensia scans your Linux servers and tells you exactly which ones are running vulnerable versions — including CVE-2026-59821. Free for 1 server.

Get started free