---
id: "CVE-2026-115352"
kind: "vulnerability"
aliases: ["CVE-2026-115352"]
title: "Mistune: inline link parser lets small markdown inputs burn CPU"
severity: "high"
cvss: 8.7
epss: 0.00359
exploited: false
status: "open"
category: "package"
products: []
ecosystems: []
tags: ["markdown-rendering", "mistune", "package", "python", "web-apps"]
not_affected_if: ["mistune-3.3.0-or-later", "no-user-supplied-markdown"]
published_at: "2026-10-22T19:17:41+00:00"
first_seen_at: "2026-10-23T21:57:31+00:00"
updated_at: "2026-10-24T07:15:51+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-115352", "first_seen_at": "2026-10-23T21:57:31+00:00"}, {"source": "epss", "url": "https://api.first.org/data/v1/epss?cve=CVE-2026-115352", "first_seen_at": "2026-10-24T07:15:51+00:00"}]
---

# Mistune: inline link parser lets small markdown inputs burn CPU

Mistune versions before 3.3.0 have a quadratic-time inline link parser. An unauthenticated attacker can send a short markdown payload of many unmatched '[' characters and tie up rendering workers for 30+ seconds per 16 KB input.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

Check the installed mistune version, for example with pip show mistune or by reading the lockfile; anything below 3.3.0 is affected.

## Fix

Upgrade mistune to 3.3.0 or later.
If upgrading is not possible, limit input size before rendering markdown.

## Timeline

- 2026-10-23T21:57:31+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-115352
- 2026-10-24T07:15:51+00:00 epss https://api.first.org/data/v1/epss?cve=CVE-2026-115352

## From the sources

Mistune before 3.3.0 contains an algorithmic complexity vulnerability in the inline link parser that allows unauthenticated attackers to exhaust CPU by submitting long runs of unmatched '[' characters. Attackers can send small markdown payloads that trigger O(N^2) parse_link_text and parse_link_label probing, occupying rendering workers for 30+ seconds per 16 KB input.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_