Fact-check findings

The reviewer could not reconcile the claim with the evidence it cited. This is a lead, not a ruling that the claim is false.

1. #7[disputed][awaiting moderator]in section: Relationship to Broader Malicious Campaign

   “The Zscaler blog post was published July 2, 2026, and updated August 5, 2026.”

   reviewerThe Zscaler blog post was published July 2, 2026, and updated August 5, 2026.The page's own archive_url for this source shows a Wayback capture timestamp of 2026-08-04T08:00:25Z — almost certainly the actual origin of the 'updated August 5, 2026' claim, i.e. an archive-capture date mistaken for an editorial update date. No independent source confirms Zscaler revised the post.
   Proposed correction (not yet applied)

   The Zscaler blog post was published July 2, 2026; no subsequent update to the post has been publicly documented.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
2. #8[disputed][awaiting moderator]in the timeline

   “Zscaler ThreatLabz updates its original blog post on the indirect prompt injection campaigns. No confirmed takedown of debank[.]auction is noted in available sources as of this date.”

   reviewerZscaler ThreatLabz updates its original blog post on August 5, 2026; no confirmed takedown of debank[.]auction is noted as of that date.This timeline entry asserts a discrete event (an August 5 update) that I could not confirm occurred at all; the 2026-08-05 date field itself is consequently unsupported. I have not proposed a replacement date because no source documents this event happening on any date — the entry's premise, not just its date, is unconfirmed, so the judge/corrector may prefer to remove this entry entirely rather than re-date it.
   Proposed correction (not yet applied)

   No subsequent update to Zscaler ThreatLabz's original July 2026 blog post on the indirect prompt injection campaigns has been publicly confirmed; no confirmed takedown of debank[.]auction is noted in available sources.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
3. #10[disputed][awaiting moderator]in section: Broader Phishing Ecosystem Targeting DeBank Users

   “Additionally, security researchers have noted that crypto drainers more broadly rely on tools like DeBank to enumerate a target wallet's holdings (tokens, NFTs, liquidity positions) before executing asset sweeps — meaning DeBank's own data layer is exploited by attackers to maximize theft yield, independent of any impersonation of the platform itself.”

   reviewerSecurity researchers have noted that crypto drainers more broadly rely on tools like DeBank to enumerate a target wallet's holdings before executing asset sweeps.The cited source does not support the specific claim attributed to it: it neither names DeBank nor describes the pre-attack enumeration mechanism the page asserts. This appears to be an invented attribution layered onto a real but unrelated source.
   Proposed correction (not yet applied)

   Additionally, security researchers have documented that crypto drainers scan a victim's wallet for valuable tokens and NFTs after a malicious transaction has been approved, in order to maximize theft yield; this mechanism has not been documented as involving DeBank's data specifically in publicly available reporting.
   citedwww.group-ib.com/resources/knowledge-hub…wallet-drainers/

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

1. #1[partially supported][awaiting moderator]in section: Entity Overview

   “DeBank (debank.com) is a legitimate Web3 portfolio tracker launched in 2018, supporting 100+ EVM-compatible blockchains.”

   reviewerDeBank (debank.com) is a legitimate Web3 portfolio tracker launched in 2018, supporting 100+ EVM-compatible blockchains, non-custodial, with a social layer called DeBank Stream.The underlying facts (2018 founding, non-custodial design, DeBank Stream social feature) are independently corroborated elsewhere, but the single source cited on the page for this section does not actually support any of these specifics — it is a generic 2026 review that never mentions a founding year, custody model, chain count, or the Stream feature by name. This is a citation-support problem, not a factual error.
   citedcryptoadventure.com/debank-review-2026-d…social-features/
2. #9[partially supported][awaiting moderator]in section: Broader Phishing Ecosystem Targeting DeBank Users

   “The domain debank[.]com-api-v2-wallet-app[.]com has been flagged by PhishDestroy as an active phishing domain associated with a crypto drainer, designed to trick users into revealing wallet credentials or authorizing asset transfers.”

   reviewerThe domain debank[.]com-api-v2-wallet-app[.]com has been flagged by PhishDestroy as an active phishing domain associated with a crypto drainer, designed to trick users into revealing wallet credentials or authorizing asset transfers.The domain is real and is logged by PhishDestroy, but the page overstates the source's actual content — PhishDestroy does not characterize the listing as an active phishing/drainer classification.
   citedphishdestroy.io/domain/debank.com-api-v2-wallet-app.com/

The cited evidence supports the claim as written.

1. #2[confirmed][no action needed]in section: Typosquatting Infrastructure

   “Structured data markup (JSON-LD) falsely identifies the site as a SoftwareApplication named 'DeBank' and misattributes debank.com as its publisher — a technique designed to pass legitimacy signals to machine consumers such as search engines and AI web-browsing agents.”

   reviewerdebank[.]auction is a combosquatting domain using keyword-stuffed metadata, fake Open Graph/X card metadata, and JSON-LD structured data misattributing debank.com as publisher, as documented by Zscaler ThreatLabz.Confirmed directly against the Zscaler ThreatLabz post.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
2. #3[confirmed][no action needed]in section: Indirect Prompt Injection Technique

   “A particularly sophisticated element: the injected prompt reportedly instructs the model to avoid using the word 'Auction' when describing the site, minimizing the signal that would alert a human reviewer to the domain's suspicious TLD.”

   reviewerThe injected prompt instructs the model to avoid using the word 'Auction' when describing the site, to minimize signal that would alert a human reviewer.Confirmed.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
3. #4[confirmed][no action needed]in section: AI Agent Vulnerability Testing

   “In the context of the DeBank impersonation campaign, two models — OpenAI's GPT-5.4 and Anthropic's Claude Sonnet 4.5 — were reported to have incorrectly rated the fraudulent site as legitimate.”

   reviewerIn testing against 26 LLMs, GPT-5.4 and Claude Sonnet 4.5 incorrectly rated debank[.]auction as legitimate, but only when not given a trusted reference to the genuine debank.com; none of the 26 models were deceived when the legitimate site was available for comparison.Confirmed against both the primary research and independent press coverage.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agentswww.securityweek.com/prompt-injection-at…crypto-payments/
4. #5[confirmed][no action needed]in section: Relationship to Broader Malicious Campaign

   “The companion campaign impersonated a legitimate Python library ('requests-secure-v2'), embedding hidden payment demands in schema markup and concealed HTML instructing AI agents to resolve fabricated errors by sending a $3 cryptocurrency payment to a hardcoded attacker wallet address: 0x691bc3793205e574fa7b4aa068e62c0e470ad267.”

   reviewerA companion campaign impersonating a Python library (requests-secure-v2) caused four of 26 tested LLMs — including versions of Llama and Gemini — to execute a $3 cryptocurrency payment to wallet 0x691bc3793205e574fa7b4aa068e62c0e470ad267.Wallet address, dollar amount and affected model families all confirmed directly against the primary source.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
5. #6[confirmed][no action needed]in section: Relationship to Broader Malicious Campaign

   “No equivalent hardcoded wallet address has been publicly reported for the debank[.]auction campaign specifically, and the financial objective of the DeBank impersonation component — whether credential harvest, wallet draining via a connected wallet flow, or AI context poisoning for downstream manipulation — has not been confirmed in publicly available reporting as of this writing.”

   reviewerNo equivalent hardcoded wallet address has been publicly reported for the debank[.]auction campaign specifically.Confirmed; this is an accurately hedged absence-of-evidence claim.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents
6. #11[confirmed][no action needed]in section: Typosquatting as Evolving Supply Chain Threat

   “LLMs generate thousands of plausible domain variants in minutes; full campaign deployment takes under ten minutes by some estimates.”

   reviewerCoverage in Hacker News (May 2026) argued typosquatting has shifted from a UI problem to a supply-chain/AI infrastructure problem, noting LLMs can generate thousands of domain variants in minutes and full campaigns can deploy in under ten minutes.Confirmed; minor caveat about the source's promotional framing noted but does not change the verdict on the specific facts cited.
   citedthehackernews.com/2026/05/typosquatting-…ser-problem.html
7. #12[confirmed][no action needed]in section: Reported Victims and Confirmed Losses

   “No confirmed individual victims or documented financial losses have been publicly attributed to the debank[.]auction domain specifically.”

   reviewerNo confirmed individual victims or documented financial losses have been publicly attributed to the debank[.]auction domain specifically; the Zscaler research was conducted in a controlled testing environment.This is a well-hedged negative claim consistent with all available reporting.
   citedwww.zscaler.com/blogs/security-research/…argets-ai-agents