The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the use of standard package management commands for project initialization and adapter setup.
Evidence: Use of npx astro add, npm install wrangler@4, and npx vercel deploy across multiple files including references/cloudflare-adapter.md and references/templates/vercel-setup.md.
[INDIRECT_PROMPT_INJECTION]: The skill provides code templates that ingest untrusted data from external sources, creating a potential surface for indirect injection if data is processed by an LLM downstream.
Ingestion points: Data enters the context via cookies and headers in references/edge-middleware.md and references/templates/vercel-setup.md, and through database/KV results in references/templates/cloudflare-setup.md.
Boundary markers: None explicitly defined in the provided code snippets to delimit external data from instructions.
Capability inventory: The skill facilitates file/state changes via CACHE.put and DB.prepare in references/templates/cloudflare-setup.md, and project configuration through npx commands during setup.
Sanitization: The snippets lack explicit validation or sanitization logic for incoming request data (cookies/headers) beyond standard JSON.stringify for output.

astro-deployment