The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-level data including Next.js source code, API routes, and database configurations which are untrusted inputs. Malicious instructions could be embedded in these files to influence the agent's behavior during deployment or troubleshooting tasks. * Ingestion points: Next.js project files, environment variables, and data from Cloudflare bindings (D1, R2, KV) accessed in route handlers. * Boundary markers: No explicit delimiters or instructions are provided to isolate untrusted application data from the agent's primary instructions. * Capability inventory: Shell command execution via npm/npx, file system modifications for project setup, and network operations for Cloudflare deployment. * Sanitization: The skill does not specify any sanitization or validation of the project content before processing it.
[EXTERNAL_DOWNLOADS]: The skill instructs users to download and install packages from the npm registry, such as @opennextjs/cloudflare and wrangler. These are well-known, official tools for Cloudflare development and are considered safe sources.
[COMMAND_EXECUTION]: The skill involves executing various shell commands for project scaffolding, testing, and deployment (e.g., npm create cloudflare, npx opennextjs-cloudflare). These are standard developer operations required for the skill's functionality.

cloudflare-nextjs