agent
Fail
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The
index.jsfile uses thechild_process.execSyncfunction with theshell: trueoption. The variableprompt, which contains unsanitized input from command-line arguments, is interpolated directly into the shell command string:opencode run "${prompt}" 2>&1. This construction is vulnerable to shell injection; an attacker can provide a payload containing shell metacharacters like;,$(...), or backticks to execute unintended commands. - [REMOTE_CODE_EXECUTION]: The command injection vulnerability in
index.jsallows an attacker to achieve arbitrary code execution on the underlying operating system with the privileges of the process running the agent.
Recommendations
- AI detected serious security threats
Audit Metadata