[GHSA-2fpx-xrc2-7qf3] A weakness has been identified in FedML-AI FedML up to 0...

[mohammedix88]

Updates

• Affected products
• CVSS v3
• CVSS v4
• Description
• Severity
• Source code location
• Summary

Comments
I read the released fedml 0.9.6 sdist and the reported sink is unchanged there, so the "up to 0.8.9" range understates who is affected. The record also has no ecosystem/package, which prevents dependency scanners from matching it at all, and the VC:L/VI:L/VA:L impact metrics do not reflect that pickle.loads() on unauthenticated remote input yields arbitrary code execution. Specific file and line references are in the description. I am not the original reporter and add no exploit detail.

[mohammedix88]

⚠️ Correction to my own submission — please do not apply the severity change.

The diff drops the existing severity array (CVSS_V3 and CVSS_V4) and sets database_specific.severity to null. That was not intentional and is not part of what I am suggesting. It happened because I edited the CVSS vector field in the improvement form and then reverted it; the form serialised the reverted state as empty rather than restoring the original values.

Please keep the existing severity block exactly as it was:

"severity": [
  { "type": "CVSS_V3", "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L" },
  { "type": "CVSS_V4", "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X" }
],

and leave database_specific.severity as "MODERATE" unless you independently decide to change it.

What I actually intend to suggest, in priority order:

1. The affected range — the record says "up to 0.8.9", but the sink is unchanged in 0.9.6, the current PyPI release. This is the change that matters most, since a user on 0.9.6 would read the current range and conclude they are unaffected.
2. The package association — adding PyPI / fedml so dependency scanners can match the advisory at all. Today it has no package and "Affected versions: Unknown".
3. The severity — raised only as a written argument in the description for a curator to weigh, not as a field change. Please treat it as a suggestion to review, not an edit to apply.

Happy to close this and resubmit cleanly if that is easier than pruning the diff. Apologies for the noise.

— Mohammad Adnan (CyStack)