The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill documents the ScriptOnce component, which allows for the generation and execution of JavaScript at runtime during the hydration process. While this involves dynamic script assembly, the examples provided use static templates for legitimate purposes like theme management.
[INDIRECT_PROMPT_INJECTION]: The skill describes a pattern for injecting external data into the document head.
Ingestion points: Untrusted data enters the agent context via loaderData in src/routes/posts/$postId.tsx, which is then used to populate the head route option.
Boundary markers: There are no explicit boundary markers or delimiters shown in the examples to isolate loaderData from the rest of the document metadata.
Capability inventory: The skill includes network operation capabilities (via the fetchPost call in loaders) and client-side script execution (via the ScriptOnce component).
Sanitization: No explicit sanitization or escaping logic is demonstrated for the data interpolated into the title or meta tags.

ssr