Tools AI Risk Radar ai-incident-0173
Incident record
Manus resolves credential exposure through an encoded email
- Severity
- High
- Status
- Patched
- Type
- Prompt Injection
- Target
- Manus with a connected email account
- Actor
- researcher
What happened
Salt demonstrated an email that caused Manus to execute encoded content when a user asked the connected agent to retrieve the message. The test exposed credentials available inside the agent’s cloud sandbox.
Salt says the issue was resolved and its demonstration no longer works. The report concerns connected-service credentials in a controlled test; it provides no fixed release number or monetary loss.
Sources
- Primary disclosuresalt.security/blog/how-we-hijacked-an-ai-agent-with-a-single…
One record from the AI Risk Radar, maintained by the Security Desk.Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 13 October 2026