Cases › security breach › LR-2026-0310

Grok and Bankr prompt-injected via Morse code to transfer DRB tokens; 14 Bankr wallets later drained

Preliminary Why: Publisher not yet named · subject to change

Developing: a major or severe case captured in the last 72 hours. Later changes are logged on the corrections page.

Status: action reported, no fix

Severity major

Reported 7 May 2026

No source states: loss or records, developer and model, agent tool, could do, deployer, industry.

Severity major: Test met: many third parties (a stated count of 14 wallets, two or more) affected, and a live production system compromised

Bankr is a crypto trading agent on X that automatically creates a linked wallet for each X account that interacts with it, including Grok's. An attacker first sent a Bankr Club Membership NFT to the Grok-controlled wallet, which the source says granted Executive permissions. The attacker then asked Grok to translate a Morse code message, and the decoded output, an instruction to send 3 billion DRB tokens, was treated by Bankrbot as an authenticated command. The agent executed it, and the tokens were transferred and immediately liquidated. Bankr later said an attacker gained access to at least 14 wallets and disabled transactions while it investigated. About 80% of the funds were eventually returned.

What happened

1 · What went wrongBankrbot treated the decoded output as an authenticated command.

2 · What it reached or did3 billion DRB tokens were transferred and immediately liquidated.

3 · How it was caughtBankr said it disabled transactions after identifying an attacker who gained access to at least 14 wallets.

What was done

About 80% of the funds were eventually returned, after the DRB community tracked down the attacker's identity.

Risks

AI used to run a cyber attack
Primary96 incidents
Rising: 24 vs 4.0 a week

AI system attacked or breached
Also23 incidents

Sorted by rule from the case's fields; not yet read from the sources.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Model

Safety training and refusal · Cyber and weapons safeguards · Abuse monitoring and account enforcement · Staged or restricted release · Dangerous-capability testing before release · Outside red-teaming · Content safety classifiers · Safeguards kept on in testing · Reasoning monitoring

Infrastructure / target

Known flaws patched · Multi-factor sign-in and password policy · Network segmentation · Intrusion and endpoint detection · Backups and recovery · Phishing and fraud filters

Evidence

Other parties affected 14

No source states: what the AI was doing, when it was found, how many records were exposed, the direct loss, any court case, the cause, what set it off.

Show the full account: 5 items from the primary sources

What happened, in order

2026-05 Morse-code prompt injection incident occurred.

Not dated in the sources:

Attacker first sent a Bankr Club Membership NFT to Grok's wallet.
Attacker then posted a Morse-encoded reply asking Grok to translate it.
The agent executed the decoded transfer instruction.
About 80% of the funds were eventually returned.

Show every field, with its source and grade