GHSA-6qv6-q77g-7qm6 is a critical-severity (CVSS 9.8) Deserialization of Untrusted Data vulnerability in nvflare. 1 public exploit reference exists, so weaponization risk is real. A fix is available for nvflare — see the affected versions and patch details below.
NVFLARE unsafe deserialization due to Pickle
EPSS Exploitation Probability
EPSS (Exploit Prediction Scoring System) is a daily probability model maintained by FIRST.org. It estimates the likelihood a CVE will be exploited in production environments within the next 30 days, derived from real-world threat intelligence signals.
How urgent is this, really
GHSA-6qv6-q77g-7qm6 plotted by exploitation likelihood (EPSS) against impact (CVSS). The shaded corner — EPSS 50%+ and CVSS 7.0+ — is where this CVE doesn't sit, though severity or exploitability alone can still warrant action.
Where this sits among everything scored
Of 377,166 CVEs with a current EPSS score, this one falls in the 10–50% band (highlighted). Real counts from FIRST.org, not a sample — log-scaled since the landscape is heavily right-skewed.
Real-World Exposure
nvflareReal-time download stats are indexed for npm and PyPI packages. This vulnerability affects PyPI packages — download data is not available via public APIs for these ecosystems.
Description
Impact
NVFLARE contains a vulnerability where deserialization of Untrusted Data due to Pickle usage may allow an unprivileged network attacker to cause Remote Code Execution, Denial Of Service, and Impact to both Confidentiality and Integrity.
All versions before 2.1.4 are affected.
CVSS Score = 9.8
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Patches
The patch is included in nvflare==2.1.4 This new version uses MessagePack instead of Pickle to do serialization and deserialization.
Some object serializations supported by Pickle are not supported by MessagePack. We have provided out of box support for some built-in NVFLARE objects. For object serializations unsupported by MessagePack, the user will need to convert the objects to numpy or bytes before sending over to remote machines. The list of supported object types are listed in https://github.com/NVIDIA/NVFlare/blob/2.1/nvflare/fuel/utils/fobs/README.rst
Workarounds
No workarounds available.
Additional information
Issue Found by: Oliver Sellwood (Nintorac) and Elias Hohl
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 🐍PyPI | nvflare | all versions | 2.1.4pip install --upgrade 'nvflare==2.1.4' |
Research use only. For defensive security, authorized penetration testing, and academic research only. Never execute exploit code against systems without explicit written authorization.
NVFLARE < 2.1.4 - Unsafe Deserialization due to Pickle
by Elias Hohl · Mar 25, 2023
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for nvflare, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.
Fix
Update nvflare to 2.1.4 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-6qv6-q77g-7qm6 is resolved across your whole dependency graph.
Workarounds
If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.
How O3 protects you
O3 Security's impact-aware SCA analyses which vulnerable code paths your application actually calls, so a match like GHSA-6qv6-q77g-7qm6 can be triaged on real exposure rather than presence alone.
Tailored to GHSA-6qv6-q77g-7qm6. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.
Frequently Asked Questions
Is GHSA-6qv6-q77g-7qm6 in your dependencies?
O3 Security finds GHSA-6qv6-q77g-7qm6 across PyPI dependencies, including transitive ones, and its impact-aware SCA ranks findings by whether your code actually calls the vulnerable path.