Prompt Injection Flaw in AWS AgentCore Harness Can Leak Credentials

Security researchers at Unit 42 found that default configurations in AWS AgentCore Harness create a gap between the agent runtime and identity management, allowing prompt injection attacks to exfiltrate credentials from agents.

Disclosed 18 September 2026 · Record updated 1 October 2026

Impact

Default configurations in AWS AgentCore Harness can allow attackers to use prompt injection to exfiltrate credentials from agents, exposing an identity/credential management gap.

Our coverage

Vulnerabilities

Unit 42 finds credential leak path in AWS AgentCore Harness

Palo Alto Networks researchers say default settings leave a gap between the agent runtime and identity management that prompt injection can exploit.

18 Sept 2026

Sources

unit42.paloaltonetworks.comhttps://unit42.paloaltonetworks.com/securing-aws-agentcore-harness-credentials