The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests manual test cases (Markdown markmap and YAML Schema) as input to generate Playwright scripts. This creates a surface for indirect prompt injection. However, the skill provides extensive safeguards, including the 'Assertion Three Questions' (断言三问) verification process and mandatory 'Clarification Rules' (提问时机), ensuring the agent validates instructions against the application's actual state before execution. The primary capabilities, such as running shell commands (npx playwright) and making network calls (apiFetch), are gated by these human-in-the-loop and verification steps.\n- [EXTERNAL_DOWNLOADS]: The skill references standard, trusted Node.js packages like playwright and @axe-core/playwright for its testing and accessibility scanning features. These dependencies are pulled from well-known official registries. Additionally, the skill actively promotes security by instructing users to manage credentials via environment variables and .env files, which are explicitly excluded from version control.

automated-e2e-testing