The Agent Skills Directory

[SAFE]: The skill provides documentation and code templates for integrating Cloudflare R2 into Workers. It correctly identifies security risks such as exposing API keys and explicitly instructs the user to use secret management tools like wrangler secret put. It also includes robust templates for CORS configuration and presigned URL generation with expiry, which are essential for secure browser-to-bucket interactions.
[INDIRECT_PROMPT_INJECTION]: The skill provides templates for handling external file uploads, which introduces a potential data ingestion surface.
Ingestion points: Data enters the environment via c.req.arrayBuffer() and c.req.formData() in templates like r2-simple-upload.ts and common-patterns.md.
Boundary markers: Not applicable, as the skill generates standalone service code rather than processing data for immediate agent consumption.
Capability inventory: The generated services utilize R2Bucket.put, R2Bucket.get, R2Bucket.delete, and R2Bucket.list capabilities.
Sanitization: Code examples demonstrate best practices for sanitization, including file type validation and the use of unique identifiers (UUIDs) for file naming to prevent overwrite attacks.

cloudflare-r2