The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs how to ingest translation data from both local files and remote API endpoints, which creates a surface for indirect prompt injection if the source content is compromised.
Ingestion points: Locale messages are loaded from local JSON files (e.g., i18n/locales/en.json) and remote API responses via the defineI18nLocale function in references/messages.md.
Boundary markers: The instructions do not specify explicit boundary markers or 'ignore' instructions for the ingested translation data.
Capability inventory: The skill uses the ingested data for rendering content via Nuxt/Vue templates and managing application routing.
Sanitization: Relies on the default auto-escaping mechanisms provided by the Vue.js template engine used in Nuxt.
[COMMAND_EXECUTION]: The skill includes instructions for executing shell commands to manage the project environment.
Evidence: npx nuxi@latest module add @nuxtjs/i18n and nuxi prepare are documented as standard setup and configuration steps in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill references the installation and use of external modules and tools from the Nuxt ecosystem.
Evidence: Fetches and installs the @nuxtjs/i18n module using the Nuxt CLI (nuxi) as described in SKILL.md.

nuxt-i18n