orchestrator-cli

SUSPICIOUS: the skill is internally coherent for task orchestration and uses official GitHub data flows, but its footprint is high-risk because it automates multiple external CLIs in dangerous approval-bypass modes, can authenticate to GitHub, and can perform unattended repo/control-plane actions. This looks like a legitimate but powerful orchestration skill, not confirmed malware.

No direct malicious exfiltration/obfuscation/backdoor behavior is evident in this fragment. The code implements a local TCP RPC supervisor with token-based authorization and runtime file/log handling, plus subprocess and process-termination capabilities. However, the provided snippet appears corrupted/incomplete (notably a large block of unrelated text inside an SQL f-string), and the most security-critical implementation (SupervisorService.start_process/send_prompt) is missing. Because 'start' is likely to execute user-supplied commands, the module can be high-impact if the RPC authorization is bypassed or the runtime directory is accessible, but there is insufficient evidence here to conclude intentional malware.