self-test
Warn
Audited by Socket on Mar 18, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill's GitHub testing purpose broadly matches its actions, but it requires a live PAT, performs real remote mutations, and forwards credentials to a gh-manager binary whose provenance is not established in the skill. No clear malicious exfiltration endpoint is shown, but the combination of unverifiable executable + credential use + autonomous remote writes makes the security risk high.
Confidence: 86%Severity: 82%
Audit Metadata