The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to manage the testing lifecycle. Specifically, SKILL.md uses k6 run and k6 inspect for script validation. It also uses script (a standard utility) to wrap CLI documentation calls to simulate a TTY environment for non-interactive execution.
[EXTERNAL_DOWNLOADS]: SETUP.md provides instructions for the agent to install xk6 via go install and build a documentation extension from github.com/grafana/xk6-subcommand-docs. These actions target official Grafana and k6 repositories, which are trusted sources for the intended functionality.
[REMOTE_CODE_EXECUTION]: Several examples (e.g., examples/browser.js, examples/functional.js) import external libraries from jslib.k6.io. This is the standard, official content delivery network for k6 helper libraries and is expected behavior for the tool.
[DATA_EXFILTRATION]: examples/cloud.js demonstrates how to stream test results to Grafana Cloud. This is a core feature of the k6 ecosystem for result storage and visualization, and uses user-provided API tokens for authentication.
[INDIRECT_PROMPT_INJECTION]: The skill involves reading external data from documentation sources (either via the k6 x docs CLI or from grafana.com). While this creates a surface for indirect injection, the sources are official vendor platforms, and the skill includes guidance on how to parse and utilize this information safely within the script generation workflow.

k6