The Agent Skills Directory

[SAFE]: The skill implements robust defenses against indirect prompt injection. The SKILL.md file contains mandatory rules that override third-party content, and the lib/utils.js file includes a sanitizeUntrusted function specifically designed to strip role-play tags and instructions found in scraped data before they are processed by the agent.
[SAFE]: API credential management follows security best practices. The instructions guide the user to store secrets in a local .env file, which is loaded via the native Node.js --env-file flag. The skill explicitly forbids the AI from echoing or displaying these keys in the chat.
[SAFE]: Network operations are restricted to the author's verified domains (openwebninja.com) and the RapidAPI proxy platform. The code does not perform unauthorized data exfiltration or access sensitive local files like SSH keys or AWS credentials.
[SAFE]: The skill's implementation is transparent and relies on standard Node.js built-in modules. No obfuscation, persistence mechanisms, or dynamic code execution from untrusted sources were detected.

openwebninja