claude-peers-mcp
Warn
Audited by Socket on Mar 22, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the skill's local peer-messaging purpose is plausible, but it achieves it by injecting peer-controlled content directly into Claude sessions and instructing users to disable normal permission safeguards. The repo/publisher mismatch and optional external summary generation add trust concerns, but there is no clear evidence of credential harvesting or outright malware.
Confidence: 82%Severity: 78%
Audit Metadata