herdr-goal
Warn
Audited by Socket on Aug 26, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the skill is purpose-aligned as a Herdr orchestrator, but it grants broad autonomous control to spawned agents using `--dangerously-skip-permissions`, processes untrusted agent/output content, and can drive real repo and deploy actions. The main concern is high-trust automation and transitive control surfaces, not confirmed malware or credential theft.
Confidence: 89%Severity: 74%
Audit Metadata