The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides patterns for mocking Inter-Process Communication (IPC) calls using mockIPC, which constitutes a surface for indirect prompt injection.
Ingestion points: Data enters the agent context through the args parameter in mockIPC handlers (found in SKILL.md and references/methods.md), which are populated by frontend invoke() calls.
Boundary markers: The skill emphasizes the use of clearMocks() in afterEach() to isolate test state, but it does not provide explicit instructions or markers to delimit untrusted data from the frontend.
Capability inventory: The skill facilitates execution of Rust commands, event emission/listening, and shell command execution via cargo test and tauri-driver.
Sanitization: There is no mention of sanitizing or validating the arguments received from the IPC mocks before they are processed by the test logic.
[EXTERNAL_DOWNLOADS]: The skill references the installation of external testing utilities.
Evidence: It provides instructions to install tauri-driver using the command cargo install tauri-driver to facilitate E2E testing.
[DATA_EXFILTRATION]: The skill includes examples of network operations to external services for testing purposes.
Evidence: Example 9 in SKILL.md and Example 7 in references/examples.md demonstrate making network requests to httpbin.org and other external URLs using the reqwest and url libraries.
[COMMAND_EXECUTION]: The skill instructs the agent or user to execute various shell commands to build, test, and install tools.
Evidence: Includes commands such as cargo test, npm run tauri build -- --debug, and tauri-driver execution.

tauri-impl-testing