text
stringlengths
0
2.67k
[1092.76 --> 1100.28] foundation that is open security so it focuses on uh securing the supply chain you know it seems to
[1100.28 --> 1106.88] have had a lot of initial energy at least uh from google so you know we first kind of came across
[1106.88 --> 1113.26] scorecard when uh dart and flutter implemented scorecards on on their github repos and they did
[1113.26 --> 1119.54] a blog post about why they'd done that and we saw that and thought yeah that's them showing us that
[1119.54 --> 1125.32] they care about security in a very sort of visual way and we immediately thought this is a good way
[1125.32 --> 1130.52] for us to communicate to our customers that we care about security so what exactly is a scorecard
[1130.52 --> 1137.46] then like is that a report card on my git commits is this a badge that i put in my banner of my readme
[1137.46 --> 1145.02] what is it so it is ultimately a badge that you put in the banner of the readme and it has a sort of
[1145.02 --> 1151.18] colored element to it so as the score gets better it goes from sort of you know red to yellow to
[1151.18 --> 1157.80] shades of green uh so once you get a score of eight or better it's a nice bright green scorecard
[1157.80 --> 1164.42] if you click on the badge that'll take you to the most recent output file from the scorecard
[1164.42 --> 1171.58] action running and so what's happening there is it's implemented as a github action every time there's a
[1171.58 --> 1178.84] merge to trunk that action is going to run and it's checking against a whole bunch of sort of
[1178.84 --> 1184.92] different areas of security so do you have known vulnerabilities are you pinning your dependencies
[1184.92 --> 1192.44] branches protected in your github config do you do static source code analysis do you do dynamic source
[1192.44 --> 1199.86] code analysis have you completed the open ssf best practices questionnaire to you know a passing grade
[1199.86 --> 1206.34] or silver or gold and so it's measuring each of those dimensions and then adding all of that
[1206.34 --> 1211.50] together into a score out of 10 you know it's really interesting i was reading this article and
[1211.50 --> 1217.66] it said gamifying like parts of like your development process you know like pipelines or different ways of
[1217.66 --> 1221.62] like writing good code and showing developers that they write good code and that they've done something
[1221.62 --> 1226.08] like i don't know if it's all because we love video games but it's proven to help people to like
[1226.08 --> 1231.36] if you gamify like your metrics you know like what colors or just anything like that they've proven
[1231.36 --> 1237.32] that people will write better code you know so it's like super interesting that i wonder if the way that
[1237.32 --> 1243.88] the scorecards have that um the little buy-in to make you want to get you know that extra five percent
[1243.88 --> 1250.92] then five more percent you know yeah it certainly does do that and you know once we started putting all
[1250.92 --> 1256.90] of our scorecards together on a table then it made me want to make sure that they were all at least
[1256.90 --> 1262.78] green but then that gets you into the whole it's about 20 percent of the effort to get 80 percent of
[1262.78 --> 1271.52] the score so it's you know relatively easy to get an eight to get from an eight to like a nine is a ton
[1271.52 --> 1277.28] more effort so you know it's a sort of pareto thing of 80 20 you know 20 percent of the effort to get 80
[1277.28 --> 1282.80] percent of the score but then it's all uphill from there because you're then putting in the other 80
[1282.80 --> 1288.26] percent of the effort to get the residual 20 of the score and it's a bit sort of law of diminishing
[1288.26 --> 1294.06] returns at some point there but honestly if people were 80 percent like more secure on a bunch of get
[1294.06 --> 1298.94] repos especially with open source because so many people are like you know then inheriting your
[1298.94 --> 1304.56] either good or bad you know what i mean so i think like even if everybody was at 80 percent
[1304.56 --> 1311.00] and just the getting it green motivated people that way like security is so optional all the time
[1311.00 --> 1315.54] until something goes wrong so i still think that would be like amazing you know it would and there
[1315.54 --> 1322.36] was a paper out a couple of months back where they looked at sort of the top 20 open source projects
[1322.36 --> 1329.58] by popularity and ran scorecard against them so i think basically none of them had a scorecard in the
[1329.58 --> 1334.06] repo but anybody can run the tool against any repo oh that's interesting because
[1334.06 --> 1340.00] if you were going to build something and you want it to look at hey is this secure enough for me to
[1340.00 --> 1344.56] you know build my application on top of it or build it into my application and the fact that you can
[1344.56 --> 1349.26] apply a scorecard to another repo that's actually really interesting that'd be like a really good
[1349.26 --> 1353.52] way to pick like what open source projects you'd want to like incorporate into something that you're
[1353.52 --> 1360.52] building you know it is i think if you're able to choose between let's say tls implementations
[1360.52 --> 1366.70] then you might very well decide you know this is the tls implementation i'm going to go for because
[1366.70 --> 1372.00] i can see that the people building this have done the things that show that they care about security as
[1372.00 --> 1377.26] opposed to you know another one where you know maybe that level of diligence isn't so transparent
[1377.26 --> 1384.18] and by making this a number or a color i'm red green colorblind so those probably wouldn't matter to me as
[1384.18 --> 1389.90] much but like seeing a number of like you got an eight out of ten it kind of boils down all the
[1389.90 --> 1395.12] complexity of hey like are you doing the best practices things are you doing the things that
[1395.12 --> 1400.18] we know you should just have this in place like you mentioned uh protecting your main branch or
[1400.18 --> 1406.24] something things that are like hey we're gonna probably avoid some of the the obvious errors in
[1406.24 --> 1413.80] how this software gets built and maintained and distributed by just doing these handful of things
[1413.80 --> 1418.80] and give you that relatively good number you know if it's like a six or a seven you're probably still
[1418.80 --> 1422.76] in the yellows but at that point you're still like oh like i you know they're doing something right it's
[1422.76 --> 1430.44] not just like an a default open github repo yeah and i think where people are implementing it you never
[1430.44 --> 1436.06] see a red badge as soon as somebody's kind of made the commitment to put a scorecard on their repo
[1436.06 --> 1442.22] they've probably also made the commitment to try and make that score at least you know decent so
[1442.22 --> 1448.48] generally you're seeing sort of you know yellow and better scorecards with scores of sort of you know
[1448.48 --> 1455.64] 6.5 or better i've not seen anybody sort of publicly putting a scorecard on something and it's you know
[1455.64 --> 1460.68] it's in an awful shape but we're gonna get better maybe later on could be a good deterrent you're just
[1460.68 --> 1465.94] like you know what this is an experiment don't use this yet it is red yeah that's
[1465.94 --> 1469.66] what i'm saying like i think that's kind of cool because when you're scoping projects and you're
[1469.66 --> 1474.34] looking for things to tools to use that would be like a great research tool and you mentioned you
[1474.34 --> 1480.36] could run the tool on any repo it's command line tool it runs in a github action what sort of
[1480.36 --> 1485.18] permission does it need on the repo itself because you said it will look for things like protected main
[1485.18 --> 1490.14] branch like is that something it's calling the github api for or is it looking somewhere else for that
[1490.14 --> 1497.06] if you're just running this on you know a bunch of random repos so it is making use of the github api
[1497.06 --> 1504.44] and you know for open source repos you can actually see that stuff so you you can see aspects of the
[1504.44 --> 1510.82] the github config yeah even though to change that config you would need to be an administrator
[1510.82 --> 1515.72] right you have read only access and that's what the scorecard needs yes and you mentioned earlier that
[1515.72 --> 1522.60] you started seeing it on flutter repos is that where it's mainly being used or where are you
[1522.60 --> 1529.00] actually seeing it being implemented in different projects and groups the energy seemed to have
[1529.00 --> 1536.44] originated from google but i think it's already spread out quite widely from there so for instance intel
[1536.44 --> 1547.16] i've now got you know a very mature open source organization the leader there um that's arun gupta he's
[1547.16 --> 1554.30] now the chair of open ssf but arun's somebody that's kind of you know moved around the industry
[1554.30 --> 1565.78] doing roles in amazon and apple and and now intel and so is recognizable industry figure and for him to
[1565.78 --> 1572.10] kind of take over as figurehead of open ssf shows that you know intel's got that commitment to
[1572.10 --> 1578.40] security but behind arun the team at intel are kind of busily working their way through
[1578.40 --> 1586.16] thousands of open source repos that they've got implementing scorecard that's just an example of
[1586.16 --> 1594.50] a large organization with a big open source presence deciding to adopt it and committing themselves to the
[1594.50 --> 1599.98] process of a implementing it and then be kind of turning the ratchet to make the scores get better
[1599.98 --> 1605.98] and you know working with their internal teams to you know change some of the tooling and culture
[1605.98 --> 1611.80] that actually lies behind achieving those scores are scorecards versioned at all it's just a kind of
[1611.80 --> 1618.92] question of like hey today you're you're a green badge you're an eight things change over time and
[1618.92 --> 1623.98] practices change and in two to three years from now i'm assuming what got you an eight today is not
[1623.98 --> 1629.64] going to get you an eight in 2027 so do they version that at all because is there a way to like notice
[1629.64 --> 1635.18] like oh which spec are you applying to because this is all to give me some assurances that you're doing
[1635.18 --> 1641.76] the right thing but that maintenance does take time so being an open source project itself scorecard is
[1641.76 --> 1648.96] constantly improving and you know one of the side effects of that is you know your eight today is not
[1648.96 --> 1653.92] guaranteed to be an eight tomorrow because the things that are being measured may move under your
[1653.92 --> 1661.52] feet so new things might be introduced hopefully it means that more people are kind of able to access
[1661.52 --> 1667.24] it so you know let's say it's measuring static source code analysis tools of course all of the popular
[1667.24 --> 1672.32] open source analysis tools are already there but there's going to be a bunch of other tools that
[1672.32 --> 1677.42] you know people have embraced and maybe that apply to languages and frameworks that aren't quite so
[1677.42 --> 1683.52] mainstream and so people are going to have the opportunity to add those in and that shouldn't
[1683.52 --> 1690.12] really affect the score of of everybody else whereas there's there's other cases of sort of new areas of
[1690.12 --> 1699.54] concern being brought in and you know that's kind of adding to the denominator everybody now has got a
[1699.54 --> 1706.36] total score that's being divided by something to get it to a score out of 10 and the divisor just got bigger
[1706.36 --> 1712.60] because there's there's more things being measured that's challenging us all i think to you know be