claude-code-expert

Overall, the skill appears to be a coherent, domain-specific toolkit for maximizing Claude Code productivity through configuration, hooks, MCPs, and sub-agents. The installed source is an official registry path, and data flows are largely local with standard API usage. The main concerns are the potential for autonomous actions and the dangerous-permissions flag, which could enable unintended actions if not properly governed. Given the combination of powerful capabilities and potentially dangerous controls, this skill is best categorized as SUSPICIOUS to MEDIUM risk pending stronger per-action governance and explicit user-approval prompts for high-risk autonomous steps.