subagent-driven-development
Fail
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: HIGHPRIVILEGE_ESCALATIONCOMMAND_EXECUTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill provides a launch pattern that includes the
--permission-mode bypassPermissionsflag when executing subagents. - This instruction explicitly attempts to override the agent platform's permission system, which is a core security guardrail.
- By bypassing these permissions, subagents can perform file system modifications, network operations, or other restricted tasks without the required user authorization or safety filtering.
Recommendations
- AI detected serious security threats
Audit Metadata