Fact-check findings

No source the reviewer could reach confirms or contradicts the claim.

1. #1[unverifiable][awaiting moderator]in section: Incident Overview

   “On the evening of September 19, 2026 (first exploit transaction recorded at 20:21 UTC)”

   reviewerFirst exploit transaction was recorded at 20:21 UTC on September 19, 2026Evening of Sept 19 and the 29-minute gap to the NTX mint are supported; the exact 20:21 UTC time was not found in the Metaverse Post article. Derived intervals (10h, 16.5h) depend on it.
   citedmpost.io/key-compromise-behind-fetch-ai-…osses-to-16-77m/
2. #4[unverifiable][awaiting moderator]in section: Scope of the Multi-Protocol Attack

   “The incident was reported by Metaverse Post to be the ninth major bridge exploit of 2026, pushing cumulative bridge-related losses for the year past $345 million.”

   reviewerMetaverse Post called this the ninth major bridge exploit of 2026, pushing losses past $345 millionThe cited Metaverse Post article contains no ninth-exploit or $345M statement. It may come from another page, but that was not found.
   citedmpost.io/key-compromise-behind-fetch-ai-…osses-to-16-77m/

A cited source no longer resolves or no longer says what the page attributes to it.

1. #8[link rot][awaiting moderator]in the cited sources

   “https://web.archive.org/web/20260922203413/https://blockonomi.com/xrp-price-eyes-2-target-as-whales-accumulate-1-54-billion-xrp/”

   reviewerCited archive for the Blockonomi sourceThe archive_url points to an unrelated Blockonomi XRP article, so it does not preserve the cited FET/SingularityNET page. I did not fetch it, and the mismatch is visible from the URL text alone. No replacement archive was found.
   citedblockonomi.com/fet-price-tanks-after-sin…conversion-pause

The cited evidence supports part of the claim but not all of it.

1. #3[partially supported][awaiting moderator]in section: Root Cause: Key Custody Failure

   “Bitquery's on-chain monitoring reported as of September 21, 2026, that the compromised conversion authorizer had not been rotated and the stolen NuNet minter role had not been revoked, meaning both attack vectors remained technically viable more than 48 hours after the initial exploit.”

   reviewerBitquery reported keys unrotated more than 48 hours after the exploitNon-rotation is supported; the Bitquery attribution and the '48 hours' figure are not in the cited article. A Sept 19 evening exploit to a Sept 21 report is also about 48 hours or less, so 'more than 48 hours' is doubtful.
   citedmpost.io/key-compromise-behind-fetch-ai-…osses-to-16-77m/
2. #7[partially supported][awaiting moderator]in section: Incident Overview

   “The attacker had already realized approximately $2.25 million through token sales by September 21.”

   reviewerAttacker realized about $2.25 million from token sales by September 21Other reporting describes about $2.25M as realized assets and funds held on the main addresses, not necessarily proceeds of token sales. The page's 'through token sales' wording is a stretch.
   citedmpost.io/key-compromise-behind-fetch-ai-…osses-to-16-77m/

The cited evidence supports the claim as written.

1. #2[confirmed][no action needed]in section: Root Cause: Key Custody Failure

   “the conversionIn function had no per-conversion token limits, allowing the full FET liquidity pool to be drained in a single call; and the signed digest did not include the recipient address as a protected field, enabling the attacker to redirect payouts.”

   reviewerBridge signer was a nonce-zero offline key; no per-conversion limit; digest excluded recipientMatches the cited article.
   citedmpost.io/key-compromise-behind-fetch-ai-…osses-to-16-77m/
2. #5[confirmed][no action needed]in section: Scope of the Multi-Protocol Attack

   “then NTX minting (408.5 million, representing approximately 42% of NuNet's total supply)”

   reviewerNuNet NTX mint was 408.5 million, about 42% of supplyProtos rounds to 400M. 408.5M appears in other reporting, so the figure holds, but the cited Protos source does not give it exactly.
   citedprotos.com/2m-stolen-in-triple-attack-on…-singularitynet/
3. #6[confirmed][no action needed]in section: Incident Overview

   “Protos reported that the attacker also conducted a preliminary sweep of ETH and BNB from approximately 16 wallets, some of which were described as belonging to SingularityNET or NuNet staff, and extracted $289,575 in USDC from a payroll contract.”

   reviewerProtos reported a sweep of 16 wallets and $289,575 USDC drained from a payroll contract; 'widespread penetration' quoteSupported. Protos attributes these to Bitquery.
   citedprotos.com/2m-stolen-in-triple-attack-on…-singularitynet/