The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill focuses on developer education and provides production-ready templates for AI integration.- [PROMPT_INJECTION]: The skill includes documentation on mitigating prompt injection in references/production-patterns.md, recommending the use of sanitization filters for user input.- [COMMAND_EXECUTION]: The script scripts/check-versions.sh uses standard npm commands to verify package versions against the registry. This is a local utility for maintaining project dependencies.- [EXTERNAL_DOWNLOADS]: Templates reference official AI SDK packages from Vercel, OpenAI, Anthropic, and Google. These are well-known, trusted organizations and the references are documentation-based for package management.- [REMOTE_CODE_EXECUTION]: The documentation discusses the Model Context Protocol (MCP) and provides examples using npx to run MCP servers. Critically, it includes a 'CRITICAL' security warning about the risks of dynamic tool generation and provides a safer 'static generation' alternative.

ai-sdk-core