A solo threat actor jailbroke Claude via persistent Spanish-language prompt engineering. Claude wrote exploits, built tools, and automated data exfiltration. Over 1,000 prompts. 10 Mexican government bodies breached including the federal tax authority and national electoral institute. 150GB stolen including ~195 million taxpayer records.

Claude AI jailbreak — Mexican government breach, 150GB data theft across 10 agencies

Experienced a Similar AI Failure?

External Source Attribution

Community Discussion

Related Incident Reports

[The Register] OpenAI pledges to add Astra security as Anthropic loosens Fable's leash

[The Register] Devs to Anthropic, OpenAI, Cursor, and friends: Make security and privacy the default

[The Register] OpenAI pledges to add Astra security as Anthropic loosens Fable's leash