The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill guides users to install various developer tools and libraries from the NPM registry.
Dependencies include vendor-owned packages (@xyo-network/), standard testing frameworks (vitest, playwright, msw), and the associated toolchain (@ariestools/).
The documentation explicitly mentions that some packages are restricted to authenticated users, which is consistent with internal development workflows.
[COMMAND_EXECUTION]: Provides numerous command-line examples for managing the development lifecycle.
Includes starting local blockchain nodes (xl1 start), running test suites (vitest), and managing test wallets (xl1-wallet).
Automation instructions for macOS and Linux include secure handling of test credentials via the OS keychain (security or secret-tool).
[DYNAMIC_EXECUTION]: Utilizes dynamic execution environments for integration and unit testing.
Employs Vitest's browser mode and Playwright to run code in headless Chromium instances to validate browser-specific APIs (IndexedDB, postMessage).
Orchestrates in-process blockchain actors for deterministic testing within the Node.js runtime.
[INDIRECT_PROMPT_INJECTION]: The skill manages a significant attack surface by ingesting data from external blockchain and datalake sources.
Ingestion points: Data is retrieved via connection.viewer (blockchain state) and RestDataLakeViewer (off-chain payload bodies).
Boundary markers: The dApp Definition of Done (DoD) requires the use of explicit schema-based validation.
Capability inventory: The skill enables transaction submission (addPayloadsToChain) and storage operations (S3 PutObjectCommand).
Sanitization: The instructions mandate a 'Zod-first' pattern, requiring all incoming data to be filtered through Zod-factory guards (isXxxPayload) before processing, which significantly mitigates injection risks.

xl1-testing