The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines a workflow for processing Markdown-based content files within a Nuxt application. This creates an indirect prompt injection surface where instructions embedded in content files could potentially influence an AI agent reading the repository. However, this is inherent to the CMS functionality and no malicious intent was detected.
Ingestion points: Markdown, YAML, and JSON files in the content/ directory.
Boundary markers: Not explicitly defined in the provided templates.
Capability inventory: Shell scripts for project setup and deployment.
Sanitization: Relies on standard framework rendering behavior.
[COMMAND_EXECUTION]: The skill provides several bash scripts (scripts/setup-nuxt-content.sh, scripts/setup-studio.sh) that automate the configuration of a Nuxt project. These scripts create directories, write configuration files from templates, and modify existing project files using standard Unix utilities like sed and cat.
[EXTERNAL_DOWNLOADS]: The provided scripts and documentation instruct the user to install well-known development tools and libraries from the npm registry, including @nuxt/content, wrangler, and vercel. A documentation error was identified regarding a non-existent version of the zod library (^4.1.12), which will result in a failed installation attempt but does not pose a security risk.

nuxt-content