The Agent Skills Directory

[SAFE]: The skill consists entirely of documentation, API references, and code implementation samples for the TanStack Virtual ecosystem. No malicious patterns, obfuscation, persistence mechanisms, or unauthorized privilege escalation attempts were detected. The content aligns correctly with the stated purpose of providing a reference for @tanstack/react-virtual.
[EXTERNAL_DOWNLOADS]: The skill references standard package installation commands for framework-specific adapters (e.g., @tanstack/react-virtual, @tanstack/vue-virtual) and related utilities like @chenglou/pretext. These are well-known libraries in the JavaScript ecosystem and are appropriate for the skill's context.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Several code samples (e.g., samples/dynamic-measured.md, references/getting-started/chat.md) demonstrate processing lists of external data such as sentences, rows, or messages for virtualization.
Boundary markers: The samples do not include specific boundary markers or "ignore" instructions for the data being rendered.
Capability inventory: The provided code is limited to UI virtualization and rendering logic; it does not include dangerous capabilities such as file system writes, network exfiltration of ingested data, or subprocess execution.
Sanitization: Standard React rendering is used, which provides basic XSS protection, though no specific logic for filtering prompt injection in the data stream is included. Given the UI-only nature of the samples, this is considered a safe surface.

tanstack-virtual