The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several CLI tools to facilitate the documentation and development lifecycle. It executes git for branch manipulation and repository synchronization, gh for interacting with GitHub Pull Requests and the GitHub API, and npm/pnpm for managing Node.js dependencies and workspace environments.
[DYNAMIC_EXECUTION]: The skill incorporates a mandatory testing workflow where code examples found in documentation and release notes are extracted and executed locally using go run . run. This involves generating temporary scripts from content that may be provided by external contributors through Pull Requests, representing a dynamic execution surface for untrusted code.
[EXTERNAL_DOWNLOADS]: To perform visual verification of documentation, the skill installs the agent-browser npm package globally and uses it to download necessary browser binaries. These downloads occur from public registries and remote servers during the skill's setup phase.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and review external markdown files, which presents a surface for indirect prompt injection attacks.
Ingestion points: Markdown documentation files located in the k6-docs repository.
Boundary markers: While the skill uses specific HTML comments like  to identify non-executable code blocks, it does not implement comprehensive delimiters to isolate documentation text from the agent's instructional context.
Capability inventory: The agent possesses capabilities to execute shell commands, perform file system operations (write to /tmp), and interact with the web through a browser tool.
Sanitization: There is no evidence of sanitization or safety filtering applied to the content of the documentation files before the agent processes them.

k6-docs