The Agent Skills Directory

[SAFE]: The skill is designed as a read-only audit tool, with instructions explicitly stating, 'Do not modify any files.'
[SAFE]: All operations utilize standard, non-destructive shell commands such as jq, grep, find, and ls to verify configuration values in package.json, .npmrc, and other manifest files.
[SAFE]: The skill references trusted external resources, specifically official Grafana GitHub repositories and documentation from Yarn, which are consistent with the identified author and tool purpose.
[SAFE]: Fix suggestions provided in the documentation are intended for manual user application and are not executed automatically by the agent.
[SAFE]: No evidence of prompt injection, data exfiltration, obfuscation, or unauthorized privilege escalation was found.

check-npm