The Agent Skills Directory

[SAFE]: The skill serves as a technical development guide for the Convex platform. It provides standard templates and documentation for building backend modules. Analysis of the instructions and code snippets reveals no evidence of command injection, remote code execution from untrusted sources, or hardcoded credentials. All external references target official documentation and well-known services.
[INDIRECT_PROMPT_INJECTION]: The skill defines a standard surface for processing external data through component arguments, which constitutes a potential indirect injection surface.
Ingestion points: Server-side functions defined in src/component/public.ts accept user-controlled arguments via the args parameter.
Boundary markers: The documentation in SKILL.md explicitly defines a security isolation boundary, stating that components are restricted from reading parent application tables, authentication context, or environment variables.
Capability inventory: The component's database capabilities are scoped to its own private schema using ctx.db operations as shown in src/component/public.ts.
Sanitization: The skill mandates the use of Convex v validators for all public API parameters to enforce strict type checking and sanitize inputs at the entry point.

convex-component-authoring