GHSA-PHMP-HRW4-R36R

Vulnerability from github – Published: 2026-11-19 22:44 – Updated: 2026-11-19 22:44
VLAI
Summary
Mistune: Quadratic-time DoS in parse_link_href via unbounded parenthesis scan on empty-label links
Details

Resolution

Fixed in Mistune 3.3.4. Upgrade to 3.3.4 or later.

  • Affected versions: < 3.3.4
  • First patched version: 3.3.4
  • Fix: https://github.com/lepture/mistune/compare/v3.3.3...v3.3.4

The reported payload was reproduced against the 3.3.3 release tag and verified to scale approximately linearly against the 3.3.4 release tag.

Original report

The following report describes the issue before the fix.

Summary

parse_link_href in mistune performs an unbounded parenthesis-balancing scan that runs to end-of-document on each unmatched (. Because the empty label [] bypasses the no_close_bracket_before high-water-mark guard, a single small document of the form "[](" * N causes O(N) openers to each drive an O(N) scan, producing O(N²) CPU on the default mistune.html() renderer with no plugins.

Root Cause

parse_link matches the empty label [] (parse_link_label("[](",1) returns ('', 2)), so it reaches the ( branch and calls parse_link_href. That function's scan loop (helpers.py:143-159) advances past characters and increments level on each (, never finding an unescaped ) at level==0, so it runs to EOF and returns None. The cursor then advances only ~2 chars to the next [](. The no_close_bracket_before high-water mark that keeps the balanced-bracket link path linear is never entered because the empty label matches before that branch; _mark_no_link_before only advances past the 2-char label, not the href region.

Impact

A single unauthenticated request rendering untrusted markdown pins a worker at 100% CPU for tens of seconds. Measured on default mistune.html, no plugins: 6KB->0.80s, 12KB->3.15s, 24KB->12.5s, 48KB->49.8s (2x input -> ~3.9x time, clean O(n²)). This is the most severe of the mistune quadratic-DoS findings.

Proof of Concept

import mistune
md = mistune.create_markdown()   # default HTML renderer, no plugins
md("[](" * 16000)                # ~48KB input stalls a worker ~50s

Attack Chain

  1. Entry: rendered markdown body -> "[](" * N (3 bytes/unit, no plugins). Guard: none on input size (grep of markdown.py/core.py = no size/time cap). Bypass proof: no global guard exists.
  2. Check: parse_link guards at inline_parser.py:141 (no_link_before), :150 (no_close_bracket_before). Guard: high-water marks that linearize balanced-bracket links. Bypass proof: empty label matches (label != None) so the :150 branch is skipped entirely; no_link_before only crawls +2/opener because parse_link_label returns ('',2).
  3. Processing: parse_link_href (helpers.py:143) scans to EOF per ( (unbalanced parens), returns None.
  4. Sink: O(N) openers x O(N) scan each -> O(N²) CPU.
  5. Impact: single ~48KB request stalls a worker ~50s (single-threaded); ~200KB approx minutes.

Bypass Evidence

  • parse_link_label("[](",1) returns ('', 2) — the empty label matches, so the balanced-bracket branch owning no_close_bracket_before is never entered.
  • git log -S "if c == \"(\":" -- helpers.py returns only refactor 5de41fb; parse_link_href has NEVER received a DoS fix (all helpers.py DoS fixes targeted parse_link_text).
  • Empirical: the [n unbalanced-bracket payload is LINEAR on HEAD (0.03->0.12s) while [](n is O(n²) (3.99x per 2x) — proving the fix does not cover this branch.

Affected Versions

<= 3.3.3 (present on latest release tag v3.3.3 and HEAD 719107f; no fix on HEAD).

Suggested Fix

Add a high-water mark for parse_link_href failed scans (mirroring no_link_before/no_close_bracket_before), or bound the parenthesis scan to a maximum length, so repeated failed href scans cannot each traverse the remaining document.


Reported by zx (Jace) — GitHub: @manus-use

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "mistune"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "3.3.4"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-97698"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-400",
      "CWE-407"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-11-19T22:44:42Z",
    "nvd_published_at": "2026-11-17T20:18:00Z",
    "severity": "HIGH"
  },
  "details": "## Resolution\n\n**Fixed in Mistune 3.3.4.** Upgrade to 3.3.4 or later.\n\n- Affected versions: `\u003c 3.3.4`\n- First patched version: `3.3.4`\n- Fix: https://github.com/lepture/mistune/compare/v3.3.3...v3.3.4\n\nThe reported payload was reproduced against the 3.3.3 release tag and verified to scale approximately linearly against the 3.3.4 release tag.\n\n## Original report\n\nThe following report describes the issue before the fix.\n\n## Summary\n`parse_link_href` in mistune performs an unbounded parenthesis-balancing scan that runs to end-of-document on each unmatched `(`. Because the empty label `[]` bypasses the `no_close_bracket_before` high-water-mark guard, a single small document of the form `\"[](\" * N` causes O(N) openers to each drive an O(N) scan, producing O(N\u00b2) CPU on the **default** `mistune.html()` renderer with no plugins.\n\n## Root Cause\n`parse_link` matches the empty label `[]` (`parse_link_label(\"[](\",1)` returns `(\u0027\u0027, 2)`), so it reaches the `(` branch and calls `parse_link_href`. That function\u0027s scan loop (helpers.py:143-159) advances past characters and increments `level` on each `(`, never finding an unescaped `)` at `level==0`, so it runs to EOF and returns `None`. The cursor then advances only ~2 chars to the next `[](`. The `no_close_bracket_before` high-water mark that keeps the balanced-bracket link path linear is never entered because the empty label matches before that branch; `_mark_no_link_before` only advances past the 2-char label, not the href region.\n\n## Impact\nA single unauthenticated request rendering untrusted markdown pins a worker at 100% CPU for tens of seconds. Measured on default `mistune.html`, no plugins: 6KB-\u003e0.80s, 12KB-\u003e3.15s, 24KB-\u003e12.5s, 48KB-\u003e49.8s (2x input -\u003e ~3.9x time, clean O(n\u00b2)). This is the most severe of the mistune quadratic-DoS findings.\n\n## Proof of Concept\n```python\nimport mistune\nmd = mistune.create_markdown()   # default HTML renderer, no plugins\nmd(\"[](\" * 16000)                # ~48KB input stalls a worker ~50s\n```\n\n## Attack Chain\n1. Entry: rendered markdown body -\u003e `\"[](\" * N` (3 bytes/unit, no plugins). Guard: none on input size (grep of markdown.py/core.py = no size/time cap). Bypass proof: no global guard exists.\n2. Check: `parse_link` guards at inline_parser.py:141 (`no_link_before`), :150 (`no_close_bracket_before`). Guard: high-water marks that linearize balanced-bracket links. Bypass proof: empty label matches (label != None) so the :150 branch is skipped entirely; `no_link_before` only crawls +2/opener because `parse_link_label` returns `(\u0027\u0027,2)`.\n3. Processing: `parse_link_href` (helpers.py:143) scans to EOF per `(` (unbalanced parens), returns `None`.\n4. Sink: O(N) openers x O(N) scan each -\u003e O(N\u00b2) CPU.\n5. Impact: single ~48KB request stalls a worker ~50s (single-threaded); ~200KB approx minutes.\n\n## Bypass Evidence\n- `parse_link_label(\"[](\",1)` returns `(\u0027\u0027, 2)` \u2014 the empty label matches, so the balanced-bracket branch owning `no_close_bracket_before` is never entered.\n- `git log -S \"if c == \\\"(\\\":\" -- helpers.py` returns only refactor `5de41fb`; `parse_link_href` has NEVER received a DoS fix (all helpers.py DoS fixes targeted `parse_link_text`).\n- Empirical: the `[`*n unbalanced-bracket payload is LINEAR on HEAD (0.03-\u003e0.12s) while `[](`*n is O(n\u00b2) (3.99x per 2x) \u2014 proving the fix does not cover this branch.\n\n## Affected Versions\n`\u003c= 3.3.3` (present on latest release tag v3.3.3 and HEAD `719107f`; no fix on HEAD).\n\n## Suggested Fix\nAdd a high-water mark for `parse_link_href` failed scans (mirroring `no_link_before`/`no_close_bracket_before`), or bound the parenthesis scan to a maximum length, so repeated failed href scans cannot each traverse the remaining document.\n\n---\nReported by **zx (Jace)** \u2014 GitHub: @manus-use",
  "id": "GHSA-phmp-hrw4-r36r",
  "modified": "2026-11-19T22:44:42Z",
  "published": "2026-11-19T22:44:42Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/lepture/mistune/security/advisories/GHSA-phmp-hrw4-r36r"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-97698"
    },
    {
      "type": "WEB",
      "url": "https://github.com/lepture/mistune/commit/b4dd3284e52ec741ed9a227db29ade9017828e90"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/lepture/mistune"
    },
    {
      "type": "WEB",
      "url": "https://github.com/lepture/mistune/releases/tag/v3.3.4"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Mistune: Quadratic-time DoS in parse_link_href via unbounded parenthesis scan on empty-label links"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Loading…

Loading…

Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.


Loading…