The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the hugo command-line utility for various tasks, including site scaffolding (hugo new), starting a development server (hugo server), and building the site for production (hugo).
[PRIVILEGE_ESCALATION]: The deployment documentation in references/deployment.md includes examples for GitHub Actions that use sudo to install the Hugo CLI (sudo dpkg -i) and the Dart Sass compiler (sudo snap install) within the CI runner environment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from various sources, including Markdown files, structured data files, and remote resources.
Ingestion points: Site content files in content/, configuration/data in data/, and remote URLs accessed via the resources.GetRemote function.
Boundary markers: The configuration guidance in references/configuration.md specifies unsafe = false for the Goldmark renderer by default, and references/templating.md documents escaping functions like safeHTML, safeJS, and safeCSS.
Capability inventory: The skill environment allows shell command execution via the Hugo CLI, network access for fetching assets, and file system modifications for site creation and updates.
Sanitization: The skill relies on Hugo's built-in template engine for context-aware escaping and requires explicit configuration to allow raw HTML.
[EXTERNAL_DOWNLOADS]: The skill references fetching the official Hugo binary from GitHub releases (github.com/gohugoio/hugo) and provides examples of retrieving data from external API endpoints during the build process using resources.GetRemote in references/asset-pipeline.md.

hugo