agent-development
Warn
Audited by Socket on Mar 4, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
The provided fragment is a benign, self-contained design guide for building Claude Code agents. It outlines best practices for delegation, tool access, memory tuning, and prompt structure without implementing executable logic or external communications. No malicious behavior, no credential handling, and no suspicious data flows are present. Overall security risk is low, and malware likelihood is negligible given the current content.
Confidence: 75%Severity: 75%
Audit Metadata