Security researchers at Salt Labs uncovered an indirect prompt-injection flaw in Manus that permitted remote code execution (RCE) in a target user’s environment. Manus AI is a rapidly growing agentic AI app which was earlier seeking funding at a $4 billion valuation. Earlier in the year, Meta’s planned $2 billion acquisition of AI agent startup Manus was formally blocked by Chinese regulators in April 2026. While Manus blocked plaintext execution requests in incoming emails, researchers successfully bypassed the app’s guardrails using JSFuck JavaScript obfuscation, triggering execution before security warnings appeared.

Full report : Salt Labs Researchers Expose Critical Indirect Injection Flaw Enabling Remote Code Execution and Credential Theft Across Integrated Cloud Accounts.