The Agent Skills Directory

[SAFE]: The skill consists of documentation and configuration examples for developers. It correctly identifies and promotes security best practices, such as leveraging environment variables for secret management instead of hardcoding credentials. It also explicitly recommends using secure communication protocols like HTTPS and WSS for all external service integrations. No malicious executable code, obfuscation, or credential theft patterns were detected.
[INDIRECT_PROMPT_INJECTION]: The skill provides guidance on integrating external services that feed data into the agent, which represents an attack surface for indirect prompt injection. This risk is inherent to the Model Context Protocol itself; the skill addresses this by encouraging structured tool schemas and input validation. No evidence of intentional vulnerability creation or malicious instructions was found.

mcp-integration