CVE-2026-105812: Code injection via unencoded configuration values during Python code generation in Bedrock AgentCore Starter Toolkit agent import
Improper control of code generation in the agent import functionality of Amazon Bedrock AgentCore Starter Toolkit before 0.3.14 might allow an authenticated same-account actor to execute arbitrary code when a user imports and runs or deploys a Bedrock Agent, via crafted configuration values incorporated into generated Python source without safe literal encoding.
To remediate this issue, users should upgrade to version 0.3.14. Because this issue persists into generated source, upgrading alone is not sufficient: agents imported with an affected version must be re-imported with version 0.3.14 or later and their local and deployed output artifacts replaced.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade to a fixed release to a version that resolves this vulnerability.
Fixed in 0.3.14 - Operational
Re-import agents imported with an affected version using version 0.3.14 or later, then replace their local and deployed output artifacts.
Event History
Frequently Asked Questions
Who can exploit this issue?
An authenticated actor in the same AWS account can supply crafted configuration values during agent import. Exploitation also requires a user to import and then run or deploy the affected Bedrock Agent.
Are agents imported before upgrading still at risk?
Yes. The unsafe values are incorporated into generated Python source, so the issue persists in local and deployed artifacts created with an affected version even after the toolkit is upgraded.
What remediation is required beyond upgrading the toolkit?
Upgrade to version 0.3.14 or later, then re-import affected agents and replace their local and deployed output artifacts. Upgrading alone does not remediate previously generated source.