The Agent Skills Directory

[METADATA_POISONING]: The SKILL.md file contains an author field set to github.com/vuejs-ai. This metadata is deceptive as it suggests an official affiliation with the Vue.js organization, while the actual provider context is rockcookies. This impersonation could cause users or agents to grant undue trust to the skill's instructions.
[INDIRECT_PROMPT_INJECTION]: The skill exposes a significant surface area for indirect injection by providing templates and code snippets for the agent to adopt.
Ingestion points: All reference files in the reference/ directory, which contain markdown-based documentation and code blocks.
Boundary markers: Absent. The skill provides no delimiters or instructions to treat the included code samples as untrusted content.
Capability inventory: The skill documents and encourages the use of shell commands for package installation (npm install), test execution (npx vitest, npx playwright), and project initialization (npm init).
Sanitization: Absent. The instructions do not specify validation or sanitization requirements for inputs handled by the suggested testing tools.
[EXTERNAL_DOWNLOADS]: The skill references and encourages downloading development tools from well-known and trusted services.
Fetches configuration and packages such as vitest, playwright, and @pinia/testing from the official NPM registry.
Links to official documentation on vuejs.org, vitest.dev, and playwright.dev for additional guidance.
[COMMAND_EXECUTION]: Instructions provide explicit commands for the user or agent to execute in the terminal, including npm install, npx vitest, and npx playwright test to manage the testing lifecycle.

vue-testing-best-practices