The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to interact with and process content from external websites using Playwright's automation capabilities. This introduces a surface for indirect prompt injection where malicious web content could attempt to influence the agent. This risk is inherent to the intended use case of web automation.
Ingestion points: Browser interaction via page.goto() and data extraction through locators as described in references/locators-and-actions.md and references/network-and-mocking.md.
Boundary markers: None identified in the provided reference guides.
Capability inventory: The skill uses npx playwright test for command execution (documented in SKILL.md) and performs file operations to store traces, screenshots, and authentication state (references/visual-testing.md, references/authentication.md).
Sanitization: Not demonstrated in the skill's reference implementations.
[EXTERNAL_DOWNLOADS]: The documentation references the installation of standard, well-known development tools and libraries from trusted registries.
Evidence: Installation commands for @playwright/test, @playwright/experimental-ct-react, and @axe-core/playwright are provided.
Evidence: The CI/CD configuration in references/configuration.md utilizes official GitHub Actions and the official Microsoft Playwright Docker image.
[COMMAND_EXECUTION]: The skill describes the use of the Playwright CLI for legitimate development tasks such as running test suites, generating code (codegen), and viewing execution traces. These commands are executed within the local development environment as part of the standard workflow.

playwright