expression-language-injection

Pass

Audited by Gen Agent Trust Hub on Sep 26, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides numerous functional payloads designed to execute arbitrary Java code on remote systems via SpEL, OGNL, and Java EL evaluators.
  • Evidence: Inclusion of payloads such as ${T(java.lang.Runtime).getRuntime().exec("id")} and reflection-based reflection bypasses in Section 2 and 4.
  • [COMMAND_EXECUTION]: Documents how to execute shell commands and capture their output using various Java framework vulnerabilities.
  • Evidence: Step-by-step instructions for exploiting CVE-2022-22947 (Spring Cloud Gateway) and Struts2 OGNL injection vectors that invoke system processes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 26, 2026, 02:02 AM