The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill involves discovering selectors on live websites, which could expose the agent to untrusted data. It addresses this by explicitly instructing the agent to ignore any text that attempts to provide instructions. This is a robust safety measure for the intended web exploration use case.- [DYNAMIC_EXECUTION]: The skill facilitates the creation and execution of local k6 JavaScript files and provides instructions for building custom k6 binaries using xk6. These processes are standard for load testing and do not introduce arbitrary or hidden execution paths.- [EXTERNAL_DOWNLOADS]: The skill utilizes utility libraries from the official k6 library host (jslib.k6.io) and references Grafana's official GitHub repositories for extensions. These sources are categorized as trusted, and their use aligns with established best practices for k6 development.

k6