sbassett added a comment.

  In T249039#6322813 <https://phabricator.wikimedia.org/T249039#6322813>, 
@Lucas_Werkmeister_WMDE wrote:
  
  > I looked at these earlier and thought they all looked like false positives
  
  Great, thanks for confirming and for your detailed analysis, with which I 
concur.  I'll change this to a risk rating of: {icon smile-o color=sky} 
**none**.
  
  > but I seem to have lost access to the paste now for some reason, so I can’t 
say for sure.
  
  This was due to some mitigations for a non-issue (T258239 
<https://phabricator.wikimedia.org/T258239>), the referenced pastes should now 
be viewable to you.
  
  In T249039#6323388 <https://phabricator.wikimedia.org/T249039#6323388>, 
@Pablo-WMDE wrote:
  
  > There recurringly are and recently were efforts to get those numbers down, 
maybe a recheck (e.g. after sha 5f1d7d106f47dbe7738efb788144d7f2fe391f39 
<https://phabricator.wikimedia.org/rEWBA5f1d7d106f47dbe7738efb788144d7f2fe391f39>)
 is all it takes to find more acceptable counts (is 0 the success criterion?).
  > This is a moving target, however. At WMDE we are in the process of finding 
a structured workflow (for the products' and the developers' sake) which 
prevents those counts climbing again. A push on T228527: Support nested 
package.json files <https://phabricator.wikimedia.org/T228527> from people with 
an official security hat would be of great help to make this happen in (ever 
more popular) monorepos.
  
  0 is of course ideal, though likely not realistic.  As noted within the 
review, outdated packages by themselves, without any additional mention of 
specific security vulnerabilities, would have a risk of: {icon check-circle 
color=green} **low**.  Per the risk acceptance chart within T249039#6309061 
<https://phabricator.wikimedia.org/T249039#6309061>, these issues can be 
addressed outside of any timeline and the risk is automatically accepted 
without managerial+ approval.  I'm also hopeful that we'll have better 
automated security monitoring in place both as stand-alone solutions and within 
CI in the near future.  Though that work is likely not to be completed for a 
while and so we try to call out such issues during manual security readiness 
reviews when prudent.
  
  > I believe this is a false positive. TinyColor (which we depend on via 
@storybook/[email protected] > [email protected] > tinycolor 1.4.1) does 
contain a copy of jquery 1.9.1 for its own demo 
<https://github.com/bgrins/TinyColor/tree/ab58ca0/demo> page, but it is not 
part of its package, and consequently not loaded in the bridge product.
  
  Ok, I'd barely call that a dev dependency then, so the risk would be: {icon 
check-circle color=green} **low**.
  
  Given the volume of issues returned by `npm audit` and `snyk test`, and that 
while such packages might not be directly deployed to wikimedia production 
hardware, they are still likely used during critical doc, test and build stages 
and I would still rate the overall risk at {icon exclamation-triangle 
color=yellow} **medium**.  This risk can be accepted by a manager (I assume 
@darthmon_wmde) and a risk plan could be as simple as committing to review 
vulnerable dependencies for security updates within 30 days (for which there 
obviously //may// or //may not// be updates.)

TASK DETAIL
  https://phabricator.wikimedia.org/T249039

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: sbassett
Cc: WMDE-leszek, sbassett, Addshore, Michael, Lucas_Werkmeister_WMDE, 
Tonina_Zhelyazkova_WMDE, Pablo-WMDE, Lydia_Pintscher, Aklapper, darthmon_wmde, 
Akuckartz, Dsharpe, DannyS712, Nandana, Lahi, Gq86, GoranSMilovanovic, QZanden, 
LawExplorer, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, Bawolff, 
Mbch331, Legoktm
_______________________________________________
Wikidata-bugs mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to