29º AVC Subcommittee Meeting | Bounded Mutable Alignment Artifact | Maker | SEED Latam

Transcript

[Music] [Music] hi everyone hello welcome already Our Guest is here good to see everyone on time hello GM GM we're going to start with our call today thank you for being on time this is our 29th ABC subc comedy meeting we have particle cope today and before we start with the topic I am going to share our

the document that we're working on about our Capt and sphere which is operational security best practices I've seen Clocky already has Shar share their document thank you so much so for everyone that would also like to contribute to this is here is that document you prefer you can comment in there you can create another document you can give us feedbacks here here all everything is going to we're going to collect all the data that H it's relevant for creating our sphere so having said that we're going to start okay these are the topics that we uh

we're discuss today we have two guests um because for this scope we are going to focus a lot in uh we're going to dive in this spell process um so we're going to try to figure out ways to improve it so let's we're going to focus on the work flow even what is spell right like the basic things um so this is the perfect moment we have Wonderland and de is ecosystem actor for today as you can see like different perspectives right from um from an outsider from a third party and from an EC consistent actor that it's involved in maker and all what

is the end game so as always uh I think this is the best moment for um if we have like any doubts or questions and maybe you have seen the recent post and a few hours also was some reply from Deer we would like to also discuss that so yes we we encourage everyone to um to have like an active participation similar like last week which was which was incredible so let's get into it I'm going to give some context first okay and then we're going to start with our guest when give them some time to introduce themsel and to discuss the

topics for today um so a few days ago wonderla post uh about this about improvements for spells inflamation process just let me see if there's anything that I am missing here from the chat hello there oh good everyone's here okay so yes Wonderland which is a team of uh I mean a group of death researchers St and scientists um investors and they're going to tell us more when they introduce themselves well um they present like this ideas in the maker forum and and they give like different

points uh and I'm going to quote this part it says like uh appro approaches to contribute to the introduction of tools that streamline the task of both spell creators and reviewers while maintaining maintaining security and minimizing the technical barriers associated with tracking and executing the process um so yes having read that also all mentioning that most of our audience I mean not everyone it's Technical and everyone is a de and everyone has um a technical background so when we start like discussing this keep that in mind that you're not talking to experts to technical experts not everyone is and so yes that way I think it's going to be

more understandable for everyone and so in the post they discusses like efforts to improve the security and efficiency and in maker uh and when it comes to like the different spells right um this I don't know these actions can range from simple transactions to significant changes in in the protocol um and so in in the post they identify like the current challenges and potential risks and the process of creating and and implementing this this

uh the Spells um so they mentioned that maker has um a meticulous checklist that creators and reviewers follow in order to ensure that spells of course are everything okay okay like they're in the right path right um but of course there have been instance U human errors uh that uh could have uh potentially compromised the protocol's Integrity I think we can all uh recall right that the last one would was a few a few months ago and that um it was shout out to clar he was the one who noticed this right like the

extra of mkr the 0.05 mkr U that and this was posted in in disc card so yes um basically they what they posting and this is just like the summary but if there is anything in particular that you would like to focus on that you have uh you would like them to uh explain in detail this is the perfect time so in the post they analyze makers H past mistakes in the execution spells they describe possible points of failure of the current process they propose a list of seven possible improvements intending to uphold or reinforce the existing this

spell creation design and our approach is not to start from scratch but to introduce tools that streamline the tasks for both crafters and reviewers enhance transparency in the process um so I think we can I'm going to share the post with everyone there you are going to see also the reply oh perfect they have share also Wonderland so you can know more about it okay this was this was like like a brief introduction but we're going to start so welcome we have a wer from Wonderland

and we have also um people from ecosis from deis shout out to them thank you for being here I going to start first with h wber since we're we're talking about Wonderland so um I'm going to give you a few times for you to introduce yourself and to tell us anything that you may want to add about uh your post in the Forum thanks please for for the introduction H let me go a little bit deeper but not too much about who's Wonderland we're a team of active Developers for cor developers like we partner with I I I wanted not to expand

this list a lot but like what where we where we are like but now in a way we are partner partner with optimism for making like the super rich maintainment will connect doing development and um incorporating like new chains and we were partnered with freeflex that we were developing like a fork of R which is high which is in a way like a a third iteration of maker DSS ecosystem no and and I I was particularly in charge of that development and maybe that's where

like I come or my vision comes from because if you see like the two protocols in way they are like quite similar and it got me to understand like many security measurements that maker D has and we are like talking about the spells that is how the actions of the like are actually onchain executed and and in a way like yep sorry um in a way these spells represent a contract that the the maker proxy will call doing what is called a delegate

call to be to run all these actions that are on this smart contract in the scope of the proxy that is like the one of the the highest like hierarchy like permissions that maker has no it's like the Le contract and it can like disable the ad if that contract calls it no it has authorization to do whatever and and these spells are executed like every two weeks so that the the proxy Maker Now like actually performs all these actions that are voted through a forum so the process

is kind of a proposal is made on the Forum then people vote for it then comes the the executive meeting or like executive voting happens and and these actions like go or not in the spell and in parallel to that there's a team actually hand coding these actions into this SM contract parallel to the executive meeting so that they know which one is approved which not H and there's as you said like a meticulous checklist that guides these actors that one a crafter one a reviewer into doing everything

according to maker standards no security standards and by the end of this process this spell this PR full request in have gets merged and then a new proposal is made an executive proposal that is like that's agree to execute all these contracts no so er we reached to this like topic I think as a way of researching like ways of governance how do they behave how do they like actually execute their their tasks and and we try to answer like the

question in the most simplest way as as we sometimes do and it's like okay how is the other people doing it no know how is maker is a since I don't know it's seven years six seven years or or has born even before like has been doing it and and and the fact that it's alive now it's and the fact that it has like such a mass means that okay this has worked like many if somebody comes from the outside and see make contracts say like can you do like some kind of other authorization system that it's not all or enough thing or can you but when you understand that it works like that it needs like the versatility of being able

to De the contract to be able to change the parameters to be able to you understand like why is the security system like so trivial in a way but um but then the focus is on on these smart contracts no and the fact that if any like malicious line of code is executed on one of these spells like the whole ecosystem can I mean not only make a out no if there's something that destroys makeer out in a way I think like no talking about any other D is going to be like completely irrelevant because we as thei

are going to cease to exist so in a way like I think we always do some trust assumptions and and working in in in software and and in IFI you have to you know you have to trust like what you get from your RPC provider is true or or that that at least you can like corroborate with another one and when you see like this SP process and and all these like meticulous checklist I I I think like they they are obviously a result of saying like Okay there was this mistake let's add this checkpoint so that this mistake is not repeated and

then there is this other or this possibility and and do it in this way and and what we've seen from the outside is that this like tiny steps are sometimes like not so well defined and and and I think something that de was told in the in the answer and was like a very interesting term to read was a tacit knowledge like I think most of these meticulous points require a lot of TD knowledge to understand about like the the previous patterns and and if you are following them and and live like lot of stuff to susceptible to human mistakes no can

be honest can be malicious but they are like human mistakes and and when you see like the whole process as a as a whole and you say like okay what's what are the trust assumptions that we we have in this process no and and they all start with a Google sheet like this is our first assumption no like the crafter and the reviewer are going to see like the the sheet that the executives have crafted and they are going to to do like everything according to that and and and if if somebody manipulates that and and nobody realizes that there's an extra line that there something that was voted or or somebody confuses like if this is approved or not then like that's our first R function then we have like the rpcs no and and

they are independently chosen by by the developers so like what's the the risk what's the risk of Google malicious manipulating a Google sheet okay I don't know I would say slow I'd say like Google has more when you talk about security and especially in in IFA you have to think about like okay what's the cost of attacking you and and what's the return that they have and thinking out maker one has to consider like this is an 8 billion organization as for now or almost um uh and when you think okay is Google going to manipulate I don't know I think Google has like bigger issues

than 8 billions maybe our RPC is going to like complot and give us like fake results okay they are independent the possibility that one is maving is low the possibility that two are misbehaving are like really low is GitHub going to like change the code ways and and not see anything okay the possibility is low there are smart contracts risks no like inheriting and like some practices that have brought some problems in in maybe more complex like code bases when when you H it from here and from this other place and there's been like cases of me

inheritance like issue like causing a lot of a lot of problems but what what we think as a vision like and I think that's General in Wonderland feeling is that like smart contracts are being are being audited our smart contract security unless we are trying to like something completely radically new they are quite okay like there are protocol that are surviving and they are not getting hacked and and they are like working okay and and we can trust the the code but something that is like happening in the ecosystem is that there are like very few upgrade processes

auditing no for example this week there was a a hack to a protocol that they did like some delegate call with an unverified check for a token transfer or something like that and and it was a very obvious B uh but what what called my attention is that this this SM contracts was the root number 406 of this possible like swap integrator no 4006 so in a way if you don't have your checks online and on Place ER it was a a time bomb it was like if it's not 406

it's going to be 407 or 450 but if you don't have like the proper testing and and procedures on what like you push into into production this is something that you can expect to happen so in a in a ro view I would say like there's not anything like very critical and I think in the in the spell repository like I think like the W has been doing a very good job like improving the testing suit and with each spell like they are getting better and and some of the proposal that we both in gith have and and in the Forum like I was it called my

attention like the the the latest spell for example I thought okay what what if we make a fight with all the transfers and then I see like the latest spell and I see that they are doing both transfers for November and for December and that's an edge case that I wouldn't have thought about you know and and it's something that because of the nature of of these spells that are like very particular to each of the situations there are like some testing techniques that are better some that are not but in a way I I have a feeling that a lot of things overall could be improved in the in the way of how how

these spells are tested how they can reduce like this do it knowledge to try to understand the whole thing that's happening how can community members like independent with lesser like technical skills can can review a process can be aware about like the the pushes that have into it how can be aware of the test that everything is on Place how can we avoid like the developers having to check Space by space if the formatting is correct and and if it follows like previous patterns how can we like standardize like some patterns and say like okay if we are like integrating a new real world asset how is the process and and try to have like a standard that

if you have to move to a standard you you need to like declare I I'm moving from the standard because this but not like every time doing something from scratch because ER there's nothing like too well designed and so so in a way I I I think like the the the the repository as a whole is a little bit susceptible to to human errors and and when you think about like the the importance that that we consider making having and and you see like okay what's the security practices that that they are implementing it's from a

perspective a little worry yet I see and and this is like something the wh was talking in the in in the foreign post and I see like they've been doing a great like Ro you know like great improvements and and and I'm happy like as an external collaborator was even like pushing PRS into GitHub and and got like amazing answers about like what's the criteria H some of the ideas were actually like taken and I think they of course like involve a lot more than than just code because then you have to check like the okay what what's the effect on

this meticulous checklist points no now how is the process but I don't know if I'm forgetting something about no I think like that could be a good presentation about who we are why we're here and on the topic I don't know if what's the spells that was actually explained in in a non- technical way no well thank you so much that was a really great presentation yes like let me know if you have like a presentation or if

you would like to share I can give you also some time just confirm it if you have if not it's it's perfect perfectly fine um no I just have like codes Snippets okay it it's okay um so from the from what you have mentioned please if anyone has any um doubts or questions this is the moment I would like now to we're gonna get back to you in a bit but I just want to also um now focus a little bit on the response on the reply from diss that basically what they uh post on on the Forum a few hours ago it's uh they

outline this steps that they took and plan to to improve the the protocol and and the the spell process and also they share like some background on how this process has evolved since um and they give like some some history right so like uh what happened before right uh with the protocol engineering C Unit resigned and they that situation left a gap um in with a new team that includes DS and Sid stream like they had to feel for that position and and the post that you are

going to see that they also like the team acknowledged that the process has room for improvement and has been actively working on on this they have improved checklist they have made uh more clear they have added missing checks for new developments and the mistakes that also they were pointed out um they have learned from them and uh they're making the process like more robust um and they mentioned even I think this is the perfect time to to discuss with both of you since in the the in the reply they also mentioned about the suggestions given for U by Wonderland

also the W were also explaining here a little bit so uh now I'm going to give the time to the team I don't know which of you from from De is going to talk but this is the moment for you to give us um a b introduction and also what would you like to highlight from the replied and also from a waiver a comment hey hello everyone um I'm F uh product manager at DWIs uh a few other members of DS are here so feel free to change with your yourself guys if you

want thank you if I like something briefly from from you from the D in general and then we can start yes yeah I can I can give a quick uh overview where we got here um and am using can go later into more technical details um basically d as you guys know uh we're an ecosystem actor um our main client is maker Dow uh we are former CES core unit uh members um and our main task or services that we offer to the our general smart contract uh implementation our main responsibility

which is shared with side stream is the implementation of spell uh code every two weeks and we share the responsibility with them where one team uh defines a crafter and a reviewer and the other team reviews as well so we keep tapped on each other make sure uh no errors fall through and we also do continuous improvements so the checklists are continuously improved the test suit is continually improved not only during the spell coding but off spell coding actually in December because there's no governance cycle that was a weak moment to finally uh put some attention there um

and trying to answer one of the comment that wiser did about for instance payments payments to delegates is something that very manual and it's uh pretty regular it happens uh every month right not in December because there's no government cycle uh that's like the edge case um and it has been manual forever like back in the protal engineering times it was manual uh it remains manual and there has been several discussions about automating it uh there are um bestest modules that is stream die and mkr by the second that that uh the owner of the best can just call best and get

their uh funds sent to their wallet so that exists there's an automation There's A Primitive already fully battle tested in the protocol is is still not used why um because the calculation of the um compensation that delegates get has been changing and I think it continues to change um and that is a bit like the main uh blocker for many automation there are other technical blockers on on adding Automation in in spellcrafting and and how the Dow operates but one of the main ones is humans and agreeing on a process that will not change um and you know payment is something that could

be agreed but it still changes and that's why it has not been automated and hopefully this whole Atlas vision about aify the core of maker will have to finally deliver those automations on chain and once those formulas are set in a stone and maybe there's a way to increase them with some inflation every now and then you can just code that up and have it immutable in smart contract and you know the numbers will not fail there will be no errors because of our contracts are deterministic same input same output we're humans we are not deterministic you give a human the same input twice and one times it can act maliciously right so I think

um when we finally have win situation is when the humans agree on a process and they agree to codify that process forever in the smart contracts maybe not forever but for a long period of time um and the the period of time needs to be long enough so the opportunity cost of focusing the development time on another project is still greater and also this I think has been one of the challenges to make here for a long time opportunity cost what collateral to Ard what project to pursue to because you know always like in businesses and even in a dow you're going to prioritize new income incremental sales than maybe maintenance or something like

that the example of maker is a bit different because in a sense we operate uh on chain money let's say you cannot make mistakes so security remains like the best you know the first uh priority but even within that generating new revenue from new collateral you know in a secure way is still sometimes uh prioritized so even if you manage to autom make payment there might be another project that delivers more value so it is a tradeoff and it's a decision that Dow needs to take and hopefully this whole Vision around aify will get us there and spells would be streamlined quite a bit I hope um yeah just just to give a bit of color on this for instance

Edge case adviser brought about the uh delegate payment in January which uh were for I think November and December or something like that um and yeah if you feel free to fill in any other uh technical details or points of view from Your Side ER okay I think you're talking about like a we were with our our contributions like trying to get in in the testing no I I do believe like the The Way spells are are crafted like the spell contract also act both as an onchain actor because it has like the V code that is going to be executed and I do think that the style of declaring

each date as a constant each thing as a constant of quing the chain log is is important you know it's a good security standard especially for this like repetitive tasks that need to be doing but and what we are we're addressing is like okay how do we test this H that all the payments are done no and and in a way our proposal was doing like a file because when when you test in solidity you you have a contract and then you have another file that has to call that contract and see that all the actions are executed properly so in the scope of adding more community community

revision or or making it clear like the test file we propose like okay what if since bests payments and and Youngs are doing like are being done in every spell or every other spell how can we like simplify the testing of this not actually the transactions being done because they are in the middle of the spell and and they may be for different reasons but how can we simplify so that any external user can see is and and understand that this spell is making all of these payments and and it's not making another one for example and and I think that this is like in a way something that may be

related with like the Craft Spells no code no that that maybe in in a future like these payments can be like Community actually crafted via a user interface and say hey we propose to make payments to this and this and this and if there is no no testing suit for for that to to be sure that that they are also not misbehaving and and and and that the transfers have be done like you all in the danger of saying like hey this user like automatically did this SP for doing this all these payments but

then a developer has to grab it has to manually test it and has to declare like hey this is good we can go so no matter how you like no code do stuff if you don't have like a a good testing process in in the end you're going to have to rely on on the human like actually saying hey this is this is okay to go yeah I fely agree I mean the who watches Watcher it's always uh yeah to Sol that that's kind of issue right because uh the thing is like a little

bit of the Technic of politics like uh the people who built original Meer protocol they had this uh very strict mindset um of pretty much favoring uh Simplicity and uh trying to uh reduce the overall complexity to the system so it could be uh more easy to reason about uh but what comes from that is that like the system that like we have today it has some clear limitations uh for instance most DS out there like for instance I think a has this like they

have two sort two different kinds of proposals uh one which is like simpler changes to the protocol and other that are considered to be like more uh core changes which have some different set of rules of being uh in order to be executed but the the governance contract of maker out is the original one like it was developed before every single one that is out there and it has some clear limitations that uh for instance doesn't allow us for easily splitting uh a spell into multiple Parts everything has to be contained in into a single contract uh and like the actual format

it has evoled a long time and uh back when we joined at CES uh the protocol engineering core unit uh had agreed to like split the the spell uh so we could work separately on a different file more or less uh the same way you are suggesting for payments uh Vice per house but uh it was actually related to managing collateral uh however like this Pro to be a little bit cumbersome especially with the amount of context that you have to keep your head when performing the

riew and of course like the process is not perfect and uh as you mentioned uh earlier in the call there is a lot of TD knowledge involved in this and I think what we've been doing and what we've been trying to do is like to sort of codify that AET knowledge into something that is rable and can actually uh help uh on board new people to the protocol being it uh an uh an outsider team uh as uh the five Wonderland or like being new team members for the existing ecosystem actors uh and this is an ongoing effort and there are kind of a tested knowledge

that is kind of hard to codify but like we we're getting there we're making improvements uh especially when it comes to the checklist like it has been completely revamped uh since we took over uh on April 2023 so uh sorry I was reading the the chat here so uh but like what one of the principles that guided uh the original team and was inherited by protocol engineering and sort of guides us is the Chesterton fence principle uh I just uh sent you a link

here explaining the rationale because uh maker engineering has a history of being uh extremely conservative uh in what it sets as uh changes in the code base so our uh authorization uh model is quite simple uh like most of the contracts has uh only one level of authorization which is the root level which pretty much allows uh anyone controlling it uh to have uh full autonomy over that specific part of the system uh some other do have like

two or three levels of permissions but uh everything is handled like that because it's really simple to reason about um and the thing is like uh the Chesterton Chesterton uh fence principle it pretty much says that like if you don't know why something was there don't go ahead and change it without understanding it first so that has a drawback of leading to some sort of alific in the process itself like uh we do that because it was done like that uh and sometimes it's hard to know the reason why something was done uh I remember like once having

a discussion back when we first joined uh we're dealing with rwa on boardings and we we realize that the rwa oracles they cannot have their price uh decreased for a specific RW collateral and nobody had seem to have a good answer of why that was the case so there are some things that end up being left to the side because otherwise like we don't know entirely what are the consequences of making that change uh but yeah the protocol changes some modules some new modules are being

developed some modules are being deprecated so we always need to make improvements in the process and yeah there were some errors like I think the the one uh regarding uh setting the the wrong value for the that sign wbtc was sort of uh more problematic uh and the the extra payment to one of the delegates that uh was the the latest one uh was also like it's not a huge problem because like uh delegates are supposed to be long-term contributors to the Dow so we can always like

uh make up to that mistake by transferring the the difference uh the later date but should point out that like in the span of four or I think yeah multic collateral die has four years or five years I never remember that actual dat but uh there were two uh errors with no uh further uh repercussions uh of course like we we don't want to make mistakes but uh this is to prove that like the process even though it seems a little bit uh weird and I agree that is not very friendly to people that are not in

the context uh but it seems to be working uh and while we want to improve it and make it easier for everybody uh we also have the mandate to keep the protocol secure so we are only comfortable of making changes we fully understand like why something was done uh in that way in the first place and what are the all possible uh implications of making that change so yeah that leads us to be a little bit slow when it comes to making changes to the spell process because uh spells are quite how can I

put this they're very sensitive because if you think about this is a bi-weekly United Boke smart contract with root access to the maker protocol uh so one mistake there uh can potentially break a protocol which is worth billions of dollars and that's not what we want to do so uh while we do agree that uh there are some lwh hanging fruits that should be addressed and they are being addressed uh we do have a list of things that we have been working and maybe uh something that we should improve is like

the transparency on that that like usually share things after we do it uh we don't have the the culture of like uh keeping posting about the progress that we make but something that we we could definitely change and uh loop uh the community in uh while the effort is being done uh part of the reason is that because most of the the discussions are pretty technical so uh if people don't have like the the right background they will probably not be able to follow the discussion but for sure like with uh things as competent as defi Wonderland uh looking at makered out we

could for sure leverage their their knowledge uh and they could for sure contribute to what we've been doing but the the in summary what I'm trying to say is that uh the the process it does require Improvement and it is currently being improved and we welcome uh any suggest from anyone but uh given like we are constrained by the time uh like the spell team currently has other responsibilities with the Dow of course when uh there is a spell and the people involved in the spellcrafting and review process uh will treat it as like maximum

priority everything else uh holds but the teams have other responsibilities so uh we we don't always have the time or the bandwidth to deal with uh every single aspect so we end up PR prioritizing what we believe uh will bring the larger benefit to the lowest cost that was it as for the chest Chester chesterton's fans I I I would say like I agree and and I think when you talk about security one has to like T steps and and be focusing on being either security neutral or security positive know one

canot like go back H in in security like trust assumptions I do believe that there are like certain CL assumtions in the code or or open to interpretation that are like this like follow previous patterns and and they are not like documented or or enforced like by lint warnings something that if somebody is like adding a new rwa collateral and they have to increase the dead ceiling they have to BU the Oracle but they forget to update maybe the collateral price for example something that is of the exec library and like having a sign that says like hey you're forgetting to

do this can be better and maybe it's not needed in that particular spell no but maybe like having to disa that sign will will bring like the reviewer some kind of a hey why why are you disabling this instead of like just not seeing something and and being like more permeable to to errors no and I understand like the the the thought behind like many of this what you say like conservative security standards but as you said like it's working and and it's working okay um I had something ah what we said

and like I think that this kind of okay every two weeks we made we make a new spell and then like some improvements are done and then the next week the next two weeks like some new spells some new improvements are done H in a way it also generates like some scattered changes in the always no and then if actually actually we avoid to make changes uh during the spell crafting uh the the window that we have uh to make improvements is in between uh spells okay interesting to know yeah this is something that is not

is the kind of tested knowledge that is not Curr documented so we could most likely this yeah I think that there was a point in the checklist that says like if some maintenance maintenance PR was done like in the last three weeks or something and there like does it change this does it change this other thing and and and I agree no and like it's it's important to to track the changes and and not like do changes on the testing when we are doing like something new because you can introduce errors in both things and and yet there are some like maybe criteria that we

will be like agree on and say like okay A a spell PR cannot change like for example this file at all no and and if there's a change in that file at all then an an alert would come and say like hey this is Chang in this file or or not even allow that to but then you have to like structure okay what do we expect as a minimum what do we expect for a particular spell testing testing suit and how are the cleanup patterns happening no because as developers what we see is changes in the code so if a malicious line was

introduced today h maybe it gets exploited two months from now and nobody nobody is going to see that that line until it it's being exploited no because it was introduced as a change right before so keeping that uh and and and making like better better testing like scripts like saying okay now we archived this this spell and we send it to another folder there is a script that actually checks if the code if the the files are exactly the same so H like improving all these uh like testing suits so that giab actually says like okay this is the pr

that archives dis spell and giab already can give you the green check of this is actually what happen and no other line was modified but then you fall in like having to trust it no like and and and that's like another security tsan that they are running okay the test and they are providing you H like a a green check when it actually that happen and and you fall in the risk of like reducing the load of what we consider like tractors that are the humans at the at the expense of TR trusting a little bit more to it have but I do

believe like there's like an analysis to make on okay what are all the trust assumptions and and be coherent with what we trust from the separate actors so that we say like so we establish if we are doing a security neutral or security positive or security negative step in saying like ah okay but this can fade in in this way but maybe we already assumed that that's not going to fail that that that that's is something like it have not be misbehaving is something core to this whole process so what I'm saying is like there there is this analysis to make on

what are we trusting and and to say if we already trusting on this maybe we can expand a little bit of the trust and and have better toolings and and not be doing any like security n step yeah for sure I mean uh in terms of the trust assumptions this is a trust process there's no other way around like it has been like this if you think about this strictly it was even worse before when maker engineering team and later the uh prodal engineering car Unit were like the single entity which was able to write and review spells and even though

like they claim they would do like red teaming and Shadow reviews and things like that uh like all those people know each other and they can pollute uh right now we do have have like two independent ecosystem actors of first we all know each other as well so in theory uh pollustion is slightly harder but uh it can also be done uh so yeah there is a trust that the developers or at least one of the developers involved in the process is honest because that that's the reason why we have uh one crafter and two reviewers because uh it would be easier to collude if where

there were less people like it could bring more but we believe that like additional rewers don't bring uh additional benefit on that so uh maybe this should be properly documented because as I said like the current governance uh the current governance module is quite uh limited and that's by Design in what it does like we could maybe move uh like as the end game evolves and we want to oy stuff in maker core for instance for Subs what we could do is leverage a more modern governance Frameworks for instance having sort some

sort of adversarial uh spell crafting in a sense that like if you publish like in order to publish a spell you need to stake some tokens uh and anyone has some time of period to challenge it with a corrected spell and the community can weigh in and like there would be slashing in place the whole cryptoeconomic thing would still uh have to be figured out but that is a possibility uh however I'm not sure we have the the bandwidth to deal with that right now because like as I said uh the endgame uh is just in the beginning and

like all teams are somehow involved in it so we have to keep the lights on with the current spells but things are going to change for sure uh another thing that I would like to to mention is that like we do have this extensive checklist uh and part of the culture of spellcrafting in maker is not skipping items or not dismissing them uh and I can say from experience that I I am someone that is usually averse to checklist I don't like them but uh when it comes to spells uh I really take it seriously and of course mistakes can be made uh but uh this is something that uh

is well thought as a way to have a back stop like the human factor as the back stop because some of the the things are really hard to automate uh the problem is that like the idea was to keep uh the protocol very open to change but that means that one can do anything the spell and that Prov to be really hard to like automate to some sort of create uh let's say factories of spells for commonly uh spell or commonly used spells uh there has been some efforts in

the past uh uh both from makeer out Engineering also from protocol engineering and we as well are discussing uh this on a more limited scope uh in the answer that we gave you we linked the the document there let me get that for you but uh we're doing that on the scope of uh emergency spells which are more critical so we are planning uh we've been discussing this for a while and we are planning to have like common emergency spell responses should be predeployment so in case this happens governance can

work without uh engineering resources being involved and uh answer to a specific emergency on a timely fashion but yeah it's an ongoing uh process and uh things uh sometimes are not as fast as we wanted to because as I said like we have the bi-weekly spell cadence and we also have the other uh duties that we uh have with the protocol so yeah we all I can say is that uh we're trying our best to improve this process and to make it more friendly because trust me like we we felt the the pain when we inherited the process and lots of things were not very

well documented uh we uh sometimes would bang our heads on the the wall trying to figure out how to do stuff uh and that brought us some learnings and uh more recently we uh on boarded some new team members and as they were uh catching uh spell duties they were would also like point out to us what were the blind spots there's currently uh there's currently a PO request open by Sid stream uh on the on the checklists in order to make it somewhat more explicit on things that must be checked checked uh all times and

things are conditional this was not clear this was some kind of uh the Tet knowledge involved uh in in the process but after this change is approved it's about to be approve should be emerged uh today and for the next spell we will most likely use this updated version then it's uh more clear but I do agree that we need to uh do better in a sense to share uh what we've been doing uh while we're doing it and not only after it's done wow guys I really didn't want to interrupt because this has been like really uh interesting discussion and to

people for example like me that are not uh experts technical experts uh we need more of this I read in the comments that probably this is are the calls from peu are going to be back so we are expecting we're looking forward to it uh I would love to keep discussing this probably we will have to we will need another protocol scope guest with both of you we don't have more time we have a call right now so thank you so much for for being here and uh this call is been recorded and I I am going to be rewatching it because there has been a

lot of insights and uh Alphas sharing here thank you so much uh to the de with Team thank you so much to W to wazer and uh that will be all for today I'm going to give you one minute guys to to say something else if you want to mention before we go so I'm going to start with uh eii ER thanks for for inviting and for making us like showing the conversation H I think that just an idea that they have it's like kind of closure is that with the with the end game and the subd I think like maker in in this time like has a little bit of the responsibility

of showing the fast forward to to all these like outs that are going to start no maybe you can like divide the risk because if some sub subow like fails is it's not a complete failure but if people will like to say like okay this is the security that we start on and and then you can like add more if you want to but that's all thanks and I think it was a very interesting conversation I hope to see to hear to part of more in the future yeah I'd like to thank uh advisor has to taken the time to contribute and try to understand the the spellcrafting

process I know it can be daunting for someone who is an outsider sometimes it's dting for people that are on the inside so yeah I really appreciate that and yeah uh we agree that we need improvements in some areas and some of the the suggestion that you made uh were like are already being worked on and some of the new things that you brought like the explicit skip for tests I think it will bring a lot of value maybe the form would be a little bit different than what you have proposed but the the intention is like all of them are all uh very good all valid so thanks for uh taking the time uh and yeah make it d i I'd say like

the the subd Dows uh to come uh even though like the core will tend to aify and be simpler uh the complexity will then lie within subd and we always need uh people with like people to to put their minds to solve the hard problems that are to come because like we're in the very frange of thei like maker is the OG defi protocol and we will be once again innovating a lot with the end game and Subs So yeah thank you very much for your contribution let's keep in touch uh it's going to be uh all good things for everyone

involved if I last words maybe e e okay sorry yeah um nothing else to add guys uh really really interesting I mean this is how humans progress sharing ideas squeezing our brains together and getting something better out so thanks a lot guys thank you everyone I'm going to share the link here on the chat so you can already watch this conversation

byebye everyone see you next week [Music]