The Agent Skills Directory

[SAFE]: The skill implements a robust 'security-first' architecture, providing explicit instructions to the agent to prioritize authentication, authorization, and data protection. The instructions in SKILL.md specifically prohibit the creation of malware, exfiltration features, or stealth behavior.
[COMMAND_EXECUTION]: The helper script scripts/check-svalbard.mjs performs local static analysis by scanning the filesystem for common security vulnerabilities (e.g., hardcoded secrets in .env, use of dangerouslySetInnerHTML, or missing security headers). The script only reads files and reports findings to the console; it does not perform any network operations or execute arbitrary external code.
[SAFE]: The PowerShell wrapper scripts/check-svalbard.ps1 uses -ExecutionPolicy Bypass to facilitate the execution of the Node.js script. This is a standard practice for development-oriented CLI tools to ensure cross-environment compatibility on Windows systems.
[EXTERNAL_DOWNLOADS]: The README provides installation instructions using npx and yarn targeting the official repository at github.com/wundercorp/svalbard-skill. These references are transparent and consistent with the stated authorship and vendor infrastructure.

svalbard