The Agent Skills Directory

[SAFE]: The skill functions as a professional development guide. It explicitly encourages best practices for security, such as requiring session secrets from the environment, using hardened cookies, and validating all data crossing trust boundaries using schemas. No evidence of prompt injection, data exfiltration, or persistence mechanisms was found.
[INDIRECT_PROMPT_INJECTION]: The skill is designed for an agent to review and build upon user-supplied application code, which is a standard surface for indirect prompt injection. The skill provides guidance on implementing validation boundaries which helps mitigate this risk at the application level.
Ingestion points: Project source code files located in app/, db/, test/, and public/ directories as defined in the project layout.
Boundary markers: The instructions prioritize "validating input at the boundary" as a core workflow step, establishing clear trust perimeters for data processing.
Capability inventory: The skill involves standard development operations including reading/writing source files and executing framework-specific CLI tools like remix test and remix doctor.
Sanitization: The skill promotes the use of remix/data-schema for runtime validation and sanitization of all external payloads, query strings, and form data.

remix