You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Abt two months ago I went and tried to cash out on a vuln I had mistakenly introduced into Candy Machine a couple years ago. My withdraw_funds function had a lack of checks and balances, and one attack vector had included messing with the candy config's authority and fake valid auth using an ages-old solana bug with passing seed authority via known bumps in CPI, further reading: #21409
Anywho, I tried to get some cash desperate in homelessness, ended up doing it successfully on devnet and ended up with a number of successful tx, i.e.:
Later that night this same code was unsuccessful on main net, suspect metaplex patched this vector in a later main net push.
when I checked about two months these devnet tx didn't exist on solana explorer, solana cli, solscan or solana.fm. they did exist on helius xray, and then I made a little angry tweet - I woke up the next morning and it was missing from xray, too. Tweet since redacted.
the above is a candy machine drain on devnet ( I think - the flow of native balance doesn't look right)
the sig 4sB56gPoffP13UNk3EZqFLpScTtjcDft1S35kFErxWviD4roNLHUWHfQJJieubBGKqxYifGLNdaKwDVKo6erdbF3 does not exist on solscan.
it does presently exist on xray, where it is an odd error that I was sometimes getting on main net, too
8. does not exist on solana.fm
9. does exist in Solana cli
10.
I gather from feedback on the previous security vuln report that this is not a security concern as reputation management doesn't count under the auspices of the definition in the policy, but it does paint a grim picture should it not be addressed.
Why are these nonaligned tx now redacted from nearly every interface? I am fully aware that I am probably uninformed about a process somewhere along the line and this is just the result of my own misinterpretation, but if people can explain this sufficiently then it does kill the vaporizing tx argument (which stems from this kind of behaviour).
Proposed Solution
Explain misinterpretation here more clearly and publicly.
The text was updated successfully, but these errors were encountered:
note: this tx 4sB56gPoffP13UNk3EZqFLpScTtjcDft1S35kFErxWviD4roNLHUWHfQJJieubBGKqxYifGLNdaKwDVKo6erdbF3 exists on both helius main net and devnet.
funny enough it is one of the main net tx that do indeed show on solscan et. al., adding to my confusion and will surely be reason to just follow up with the folks at Helius to see why this tx appears on both main net and devnet.
I still believe this is indicative of a greater issue
I've also never heard of Big Table causing issues with some tx not being saved longer or shorter than others, if anyone can elaborate
hi @t-nelson
Problem
Abt two months ago I went and tried to cash out on a vuln I had mistakenly introduced into Candy Machine a couple years ago. My withdraw_funds function had a lack of checks and balances, and one attack vector had included messing with the candy config's authority and fake valid auth using an ages-old solana bug with passing seed authority via known bumps in CPI, further reading: #21409
Anywho, I tried to get some cash desperate in homelessness, ended up doing it successfully on devnet and ended up with a number of successful tx, i.e.:
5GwvNXcCVoCkwFcYQtJNshjyqvHsiNCbFV3w9HAdGmc8KPZ1JZu4NruuJpyrp7aGTbz8naoGHra4Z3AiCZyyBde5
among many others
Later that night this same code was unsuccessful on main net, suspect metaplex patched this vector in a later main net push.
when I checked about two months these devnet tx didn't exist on solana explorer, solana cli, solscan or solana.fm. they did exist on helius xray, and then I made a little angry tweet - I woke up the next morning and it was missing from xray, too. Tweet since redacted.
Today I mucked around, if you check candy machine v1 on solscan for devnet you find nothing newer than 4 months ago
https://solscan.io/account/cndyAnrLdpjq1Ssp1z8xxDsB8dxe7u4HL5Nxi2K5WXZ?cluster=devnet
run this and you see block times as recent as Oct 11 2023, atow a week ago
Tun the following code
you get this wee lil guy:
8. does not exist on solana.fm
9. does exist in Solana cli
10.
I gather from feedback on the previous security vuln report that this is not a security concern as reputation management doesn't count under the auspices of the definition in the policy, but it does paint a grim picture should it not be addressed.
Why are these nonaligned tx now redacted from nearly every interface? I am fully aware that I am probably uninformed about a process somewhere along the line and this is just the result of my own misinterpretation, but if people can explain this sufficiently then it does kill the vaporizing tx argument (which stems from this kind of behaviour).
Proposed Solution
Explain misinterpretation here more clearly and publicly.
The text was updated successfully, but these errors were encountered: