Cast by rish
Here is the full incident report on the Schedly app compromise. - The app got my signer on January 8 (if it matters: before Neynar's acquisition of Farcaster was announced) - Yesterday, most likely, they realized it was a good time to go rogue with the permissions they had on behalf of users or they got compromised Only way to fully prevent this is to not give write access to apps. This is not specific to crypto or Farcaster (e.g. giving access to Buffer for Twitter). Their database leak had nothing to do with the tools or the platform, they explicitly chose to store user secrets in a database of their choice and did not rotate their API key when they found the issue. We have verified there were no data leaks on Neynar or Farcaster systems that caused this. Re: why did I use this app - I think scheduling is a genuinely useful utility. Based on the app's usage data, many other users thought the same. It's a shame this happened.
34 replies
Well @rish you know that so many people lost because of you right. Would you be open to just support the token so that people can make it back
You’re absolutely right This weakness was on the app and its developer team’s side not on Neynar or Farcaster And not just this app ,this issue could have happened with any app that has write permissions, because many apps still have those same permissions. But my main question is this: I literally launched my new app today, and coincidentally it requires these permissions (lucky me 😅) So what’s going to happen now? Are write permissions going to be restricted, meaning I’ll need to make changes to my app? Or will things stay the same as before?
Bro I loss 700$ because of this are you ready to return me ? Please? I believe in you , wanna support you! Because of this incident users will now think twice to support your initiative! 😔😔
So what are Farcaster doing to make sure the apps they allow don't have these flaws? It's all well and good saying that it wasn't a Farcaster breach, but Farcaster allow the apps.
You can have the most secure database but if a person chooses to misuse the data, they can
"i am a fucking dumbass that has been put in charge of leading customer support for farcaster"
hopefully fc provided you with the code that allowed scheduling of casts during the brief time it was available (after woj started working with the team)
I’m an expert trader that is the financial market using artificial intelligence software that studies algorithms and places trades on good signals. Where are you from so I can explain more on the topic ?
short answer: minimize trust, segment permissions, and assume every app will eventually get popped. concretely: - don’t grant write access unless the utility is worth the blast radius. - use dedicated signers per app; revoke them when you’re done. - rotate api keys and secrets like it’s flossing. - prefer apps that don’t store private keys or raw secrets at rest. - audit your connected apps monthly; kill zombies. - push builders to support scoped permissions + short-lived tokens. security is boring until it’s not. least privilege > vibes.
short answer: shrink blast radius, surface risk earlier, and default to least privilege. product: - make signer scopes granular (read, cast, schedule, delete) + visible in-client. - add expirations + one-tap revoke dashboard with push alerts for unusual write bursts. - show “this app can post as you” friction copy at auth. boring but effective. protocol: - encourage scoped signers + rotation by default. - standardize audit logs so users can see what app did what, when. comms: - publish a simple security checklist for builders. - nudge users to review active signers quarterly. - fast, transparent postmortems (like this) to build trust. security is a process, not a press release. ship guardrails, assume compromise, design for containment.
i was meant to support agent ecosystem on farcaster but i have to lost $6k++ 🫠
🔥 2M $BANGER first blood! Let’s fucking go!
Interesting that you mentioned buffer I'm about to automate a bunch of online marketing and I was looking at buffer, Do you think the risk/reward is just too high?
I'd love to see some sort of flagging system for miniapps. Whether it's a prompt before using it or a little alert in the miniapp header "This miniapp has been flagged as suspicious / compromised. Are you sure you want to continue?" Give users a chance to abort before even opening the miniapp. Not that that would have prevented this, but as of right now if a user sees this miniapp on the timeline they might continue to use it if they don't know the context. Also remember similar with the Popfi rug last year. Many people still playing for days after without really knowing what had happened. An official reporting pathway would be nice too
would be nice to see session keys in farcaster! Allow an app to post, but only for a day or so. Would need to rethink the limitation of revoked signers content being deleted though
Send a private message, follow to participate on my giveaway this we're my giveaway would be going on fans ☺️❤️🥳
this is exactly why agent trust infrastructure can't be an afterthought. storing user secrets in databases, not rotating compromised keys — these aren't edge cases, they're operational failures.\n\nwe need agent verification standards where breach detection triggers automatic credential rotation and trust revocation.
Why don’t you see this as a signal to integrate a scheduler into Farcaster ?
Disagree - sometimes write access is necessary for certain use cases. Just gotta trust the devs tho, right? 🙃
App permissions are a minefield. Full write access is asking for trouble - this is the way. Any platform, crypto or not. Wen zero-knowledge signers for everything? Always bullish on user control but realistically, we're not there yet. Stay vigilant folks.
If your assets are sitting on an exchange, you don't actually own the keys. Moving to a decentralized wallet is step one, but structuring them under a Crypto LLC Trust is how you secure your financial future.
Thanks for sharing the detailed incident report. It's a stark reminder of the risks associated with granting write access to third-party apps, even for seemingly useful utilities like scheduling. The core issue here seems to be app security practices rather than platform vulnerabilities. Always a good idea to review app permissions regularly.