fable-herdr-orchestration
Fail
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: HIGHCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill explicitly instructs the user and the agent to execute commands using the
--dangerously-skip-permissionsflag. This flag is designed to bypass interactive permission prompts, allowing the agent to perform file system modifications, network requests, or other sensitive operations without explicit user approval for each action. - [COMMAND_EXECUTION]: The orchestration logic involves launching multiple worker instances (
claude,codex) with safety checks disabled. This creates a multi-agent environment where tasks are decomposed and executed autonomously across different sessions, significantly increasing the risk of unvetted command execution if the 'control tower' agent is compromised or follows a malicious instruction.
Recommendations
- AI detected serious security threats
Audit Metadata