The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources, which constitutes an indirect injection surface.
Ingestion points: The instructions facilitate loading Rive assets from remote URLs (File.url, FileLoader.fromUrl) and raw byte buffers (File.decode) in SKILL.md.
Boundary markers: There are no instructions for implementing boundary markers or verifying the provenance of external Rive files.
Capability inventory: The skill uses the rive_native package which executes native C++ code via FFI and performs network requests to fetch animation assets.
Sanitization: The skill does not include steps for sanitizing or validating binary asset data before it is processed by the native runtime.
[EXTERNAL_DOWNLOADS]: The skill references and downloads components from well-known and official sources.
Evidence: Fetches package dependencies rive and rive_native from the official Flutter registry (pub.dev).
Evidence: References official documentation and assets from the rive.app domain and github.com/rive-app repositories.
[COMMAND_EXECUTION]: The skill includes a command for environment setup related to native binary management.
Evidence: dart run rive_native:setup --verbose --clean --platform macos is used to configure the native runtime environment as part of the standard development workflow.

rive-flutter