The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents how to build applications that ingest and process untrusted external data, which is an inherent attack surface for indirect prompt injection.
Ingestion points: External data enters the application context via the file system (tauri-plugin-fs), network requests (tauri-plugin-http), and IPC calls (invoke).
Boundary markers: The instructions mandate the use of Tauri 2.x Capability files to restrict access and Content Security Policy (CSP) to limit webview execution.
Capability inventory: The skill describes capabilities for file writing, shell execution, and network operations across all examples.
Sanitization: The skill relies on standard Rust serialization (serde) for IPC, but final data sanitization logic is left to the developer implementation.
[COMMAND_EXECUTION]: The skill provides instructions for using the tauri-plugin-shell to execute external processes. This is documented with critical security warnings to avoid unscoped execution and to explicitly list allowed commands and arguments.
[EXTERNAL_DOWNLOADS]: The skill references external resources for configuration and services.
It fetches a metadata schema from the author's own infrastructure on GitHub.
It references a well-known certificate authority (DigiCert) for binary timestamping configuration.
Documentation examples for update servers and APIs use safe example.com placeholders.
[PERSISTENCE]: The skill includes documentation for the tauri-plugin-autostart plugin, which enables applications to launch on system login. This is a standard feature for desktop applications and is presented as an optional plugin within the official registry.

desktop-framework-tauri