fable-herdr-orchestration
Fail
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: HIGHCOMMAND_EXECUTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The instructions in
SKILL.mdadvise the user to launch the AI agent session using the--dangerously-skip-permissionsflag. This action bypasses the environment's built-in security architecture that requires user confirmation for high-risk operations. - [COMMAND_EXECUTION]: The skill provides several shell command templates in
SKILL.mdfor launching worker agents that include flags to skip permission checks and configurations that automate approval reviews, effectively removing the human-in-the-loop safety mechanism.
Recommendations
- AI detected serious security threats
Audit Metadata