The Agent Skills Directory

[SAFE]: The skill defines instructional workflows for backend design using the Convex platform. It does not contain any obfuscated code, malicious network operations, or persistence mechanisms.
[COMMAND_EXECUTION]: The instructions include the use of npx tsc --noEmit and npx convex dev to verify code correctness. These commands use official developer tools from the platform vendor (Convex) and are intended for legitimate static analysis and deployment verification.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user product ideas to generate backend code. While this represents a standard data ingestion surface, the lack of dangerous capabilities or high-privilege operations renders this surface a low risk inherent to its function.

convex-design