fix: backport DOS via __proto__ key in merge config fix to v0.x#7388
Conversation
|
running tests locally, 1 test fails the same test fails on the v0.x branch |
There was a problem hiding this comment.
1 issue found across 3 files
Confidence score: 4/5
- Minor risk: the
__proto__tests intest/unit/core/prototypePollution.jsmay not actually exercise own-property filtering because object literals treat__proto__as prototype assignment, so the test could give a false sense of coverage. - This looks like a low-severity testing concern (4/10, low confidence), so the change appears safe to merge with minimal risk.
- Pay close attention to
test/unit/core/prototypePollution.js- ensure the__proto__filtering behavior is truly validated.
Prompt for AI agents (all issues)
Check if these issues are valid — if so, understand the root cause of each and fix them.
<file name="test/unit/core/prototypePollution.js">
<violation number="1" location="test/unit/core/prototypePollution.js:17">
P2: Tests attempt to verify `__proto__` filtering using object literals, but `__proto__` in object literals sets the prototype instead of creating an own property. As a result, the merge/filtering logic for `__proto__` is not actually exercised, leaving prototype-pollution protection untested in these cases.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
__proto__ key in merge config fix to v0.x__proto__ key in merge config fix to v0.x
|
@FeBe95 im going to merge this and will then look to release soonish. i need to fix the ci etc to actually release this and keep v0.x working correctly / better which i will all try get done this week |
|
@jasonsaayman could you share the ETA on release update on v0.x with this fix? |
|
@gbhasha Version 0.30.3 has just been released! |
|
Hi @FeBe95 , @jasonsaayman Many enterprise projects are still dependent on the 0.x line. Are there any plans to fully address this issue and clear all related vulnerabilities in the 0.x branch as well (including v0.30.3)? We’re currently seeing security scanners continue to flag the 0.x version, so clarification on the maintenance and support plan for 0.x would be very helpful. Thanks in advance. Refs:https://security.snyk.io/package/npm/axios/0.30.3
|
|
Hey @VikramSingh1994 :)
The vulnerability was fixed in version
I have never used Snyk myself, but I guess that they are showing outdated vulnerability information then. They probably automatically flagged version Note that the NPM registry (for The affected version ranges should look like this now: The GitHub advisory database allows everyone to propose changes via the "Suggest improvements for this vulnerability" link on each page. Snyk does the same via the "Found a mistake?" button. You have to have an account for this though, which I don't have. Feel free to make them aware of the fixed vulnerability in version https://security.snyk.io/vuln/SNYK-JS-AXIOS-15252993
|
|
@VikramSingh1994 if the issue still shows up in like a day or two pls reach out and i can chat with SYNK / github sec team |
|
Snyk seems to have updated their page. The H(igh) vulnerability in
@jasonsaayman They are still showing a M(edium) vulnerability in
I saw that the GitHub advisory database lists the correct version ranges, while the advisory in this repository wasn't updated. And it also says
Could you please update this as well? Maybe Snyk is taking the latter one as a source for their CVE details.
vs.
|
|
done |
|
Thanks @FeBe95 @jasonsaayman , |
|
How about define |







Backports fix for GHSA-43fc-jf86-j433 on v1.x (#7369) to v0.x.
Summary by cubic
Backports the GHSA-43fc-jf86-j433 prototype pollution fix to v0.x by filtering dangerous keys during config merging. Prevents DoS via proto/constructor/prototype and aligns v0.x with the v1.x behavior.
Description
Testing
Written for commit a615fe5. Summary will update on new commits.