# [CRIT] [GHSA / CRITICAL] CVE-2026-7304: SGLang: Unauthenticated RCE via --enable-custom-logit-processor

**Source:** GitHub Security Advisories
**Published:** 2026-05-18
**Article:** https://github.com/advisories/GHSA-36m8-w8qf-g76p

## Threat Profile

SGLang: Unauthenticated RCE via --enable-custom-logit-processor

SGLang's multimodal generation runtime is vulnerable to unauthenticated remote code execution when the --enable-custom-logit-processor option is enabled, as Python objects loaded via dill.loads() will be deserialized without validation.

Affected packages: pip:sglang (vuln >= 0.4.1.post7, <= 0.5.12).

Tracked as CVE-2026-7304.

## Indicators of Compromise (high-fidelity only)

- **CVE:** `CVE-2026-7304`

## MITRE ATT&CK Techniques

- **T1190** — Exploit Public-Facing Application

## Kill chain phases observed

_(none detected from narrative keywords)_

## Recommended hunts

### IOC-driven hunts (use shared templates)

These are standard IOC-substitution hunts — the canonical SPL and KQL live once in [`_TEMPLATES.md`](../_TEMPLATES.md), so we don't repeat the same boilerplate on every CVE / hash / network-IOC briefing.

- **Asset exposure — vulnerability matches article CVE(s)** ([template](../_TEMPLATES.md#asset-exposure)) — phase: **recon**, confidence: **High**
  - CVE(s): `CVE-2026-7304`


## Why this matters

Severity classified as **CRIT** based on: CVE present, 1 use case(s) fired, 1 technique(s) inferred. Read the full article for actor attribution, tooling details, and any defanged IOCs in the body that aren't visible in the RSS summary.