Beeper
send a message anyone will read. built by @to & @beep.eth
Latest casts
small detail, big first step for Farcon Rome, @beep.eth helped bring NFC name tags to life — letting people tap to connect IRL beeper is still early but this was our first offline experiment: taking the attention layer we’re building online and giving it a physical touchpoint tap to connect 🦖 beep beep
What we've done and what's next: • Patched the vulnerable endpoint and restored the monitor-layer sender binding • Removed all 774 fake notifications from user inboxes • Restored Beeper after patching and verification • Auditing every remaining API route for missing authentication or similar gaps • Adding regression tests for auth on all identity-bearing endpoints • Keeping heightened monitoring in place as service resumes To everyone who saw the fake notification, and to the @farcaster team for the disruption this caused — we're sorry. This should not have been possible, and the responsibility sits with us. Thank you to the people who flagged it early, to @cassie for digging into the onchain history, and to @farcaster for the public warning that helped limit the damage. If you have concerns about your account or received the fake notification, DM @beep.eth directly. https://farcaster.xyz/cassie/0x5e0c9eba
Impact: 774 fake notifications sent, reaching 6,699 users across 4 impersonated identities. 81 were opened. 0 clicked. 0 claimed. $0 moved. What was not compromised: servers, smart contracts, signing keys, backend credentials, user wallets, user funds, or any user accounts on Farcaster or elsewhere. Beeper has never handled Farcaster signer keys — we have no ability to cast on a user's behalf.
The attack originated from wallet 0x8fe3c6a86840b09e36647a9bbe4b8b6041314523 On Farcaster, that wallet is associated with @galaxysync (FID 1895288). In Beeper, a Farcaster user record tied to that identity was created ~4 hours before the attack, with displayName set to the literal string "Farcaster". The first transaction from this wallet was self-labelled "Sender Spoof Proof". The same attacker was responsible for all 774 spoofed messages across the 4 impersonated identities.
Technical root cause — a defense-in-depth regression across two layers. Layer 1 (application): POST /api/v1/quotes had been hardened with withAuth in the past; that hardening was not present in the attack-time version. The handler ran without an auth wrapper and accepted userId from the request body, passing it into the service layer as sender identity. Layer 2 (monitor): the blockchain monitor historically refused to create a Message row unless the on-chain tx's from matched one of the claimed sender's registered wallets. That check was dropped during a MultiToken consolidation refactor (~25k lines deleted) that merged three legacy handlers into one. The omission wasn't caught because later features (router-mediated sends, x402 agent payments) legitimately require decoupling on-chain payer from sender, making the absent binding appear intentional in review. With neither layer enforcing the binding, each assumed the other was.
Timeline (UTC) 03:44 — attack began (774 spoofed notifications over a short window) 04:20 — community members escalated the issue to us; Beeper was paused within the same minute 04:30 — initial public notice posted 05:15 — vulnerability patched, all 774 fake messages removed from inboxes Now — service has been restored after verification; broader audit and heightened monitoring remain in progress.
Earlier today, a vulnerability in Beeper's API was used to send 774 notifications impersonating 4 identities, including @farcaster We shut the app down immediately, removed all fake notifications from inboxes, and have now restored service after patching and verification. No Farcaster accounts, signer keys, wallets, or user funds were compromised. Full breakdown below.