The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process external data such as browser console logs, server logs, and SSE streams to diagnose issues. While this represents an ingestion surface for potentially untrusted data, the skill is purely instructional and does not provide tools for automated file writing or command execution that could be exploited by such data. The risk is consistent with standard developer debugging workflows.
[EXTERNAL_DOWNLOADS]: The skill references the vendor's official MCP server (https://mcp.copilotkit.ai/mcp) and documentation sites (docs.copilotkit.ai). These are trusted resources belonging to the skill author ('copilotkit') and are used for looking up troubleshooting information and code examples.
[COMMAND_EXECUTION]: The diagnostic steps suggest running standard development commands like npm ls and curl against local endpoints (e.g., localhost:3001) to verify connectivity. These are routine debugging operations and do not involve executing arbitrary or remote code.

copilotkit-debug