The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform standard development tasks using the shell, such as scaffolding projects with npm create cloudflare@latest, managing dependencies with npm install, and deploying code via wrangler deploy. These operations are consistent with the skill's stated purpose of helping developers build and publish Cloudflare Workers.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it accepts project descriptions from the user to determine the project structure, dependencies, and configuration bindings. This input directly influences the agent's output and actions.
Ingestion points: The project description requested in the 'Step 1: Understand the Project' section of SKILL.md.
Boundary markers: The instructions do not define delimiters or provide 'ignore embedded instructions' warnings for the data processed during scaffolding.
Capability inventory: The agent has the capability to write local files (wrangler.jsonc, package.json, etc.), install arbitrary packages via npm, and execute deployment commands that interact with a Cloudflare account.
Sanitization: There is no evidence of validation or sanitization of user input before it is interpolated into the scaffolding workflow.
[SAFE]: All external URLs and dependencies are sourced from trusted organizations and well-known services, including Cloudflare's official documentation and GitHub repositories (github.com/cloudflare/*) and the Hono framework's resources (hono.dev). These are documented neutrally and do not escalate the risk profile.

cloudflare-worker-builder