The Agent Skills Directory

[SAFE]: The skill is entirely focused on providing documentation, best practices, and code templates for Cloudflare D1. It includes a shell script for setup and TypeScript examples for database interactions.
[PROMPT_INJECTION]: No evidence of prompt injection attempts or instructions to bypass safety guardrails was found. The instructions are technical and directed at the agent's task-specific behavior for database management.
[DATA_EXFILTRATION]: No patterns of data exfiltration or hardcoded credentials were detected. The documentation explicitly advises users to use environment variables or secrets for sensitive identifiers like database_id.
[COMMAND_EXECUTION]: The provided shell script template (templates/d1-setup-migration.sh) uses standard developer tools like npx wrangler. These commands are expected within the context of managing Cloudflare Workers and D1 databases and do not represent unauthorized command execution.
[REMOTE_CODE_EXECUTION]: The skill does not download or execute untrusted remote code. It references official Cloudflare documentation and uses the official wrangler CLI via npx.
[DATA_EXPOSURE]: The skill processes user inputs (like email addresses or user IDs) to generate SQL queries. However, it mandates the use of prepared statements and parameter binding (.prepare().bind()), which is the standard industry practice to mitigate SQL injection risks.

cloudflare-d1