review
Warn
Audited by Socket on Mar 29, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The core PR-review purpose is coherent, and GitHub CLI usage is consistent and mostly benign, but the skill’s mandatory dynamic discovery and execution of all local plugin agents/skills creates a broad transitive trust chain and prompt-injection surface. Main risk is not overt exfiltration; it is overbroad autonomous orchestration of unreviewed local extensions.
Confidence: 88%Severity: 72%
Audit Metadata