The Agent Skills Directory

[COMMAND_EXECUTION]: Utilizes shell commands (find, grep) to inspect the local filesystem for existing testing infrastructure and configurations.
[COMMAND_EXECUTION]: Executes package management commands (bun add) and tool initialization (bunx playwright install) to set up the testing environment.
[EXTERNAL_DOWNLOADS]: Downloads browser binaries from official distribution channels during the Playwright setup process.
[EXTERNAL_DOWNLOADS]: Accesses external websites (playwright.dev, npmjs.com) via WebFetch or WebSearch tools to retrieve the latest version information for dependencies.
[REMOTE_CODE_EXECUTION]: Configures an MCP server to enable automated browser interactions by the agent. The skill specifies the execution of @playwright/mcp@latest via bunx. This package name is inconsistent with the official @anthropic/mcp-server-playwright package mentioned in the initialization steps, which presents a risk of executing unverified code from a potentially typosquatted or malicious package.
[COMMAND_EXECUTION]: Modifies persistent project configurations, including .mcp.json, and creates GitHub Actions workflows (.github/workflows/e2e.yml), establishing automated code execution paths in both local and CI/CD environments.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes local project data to determine its configuration state.
Ingestion points: Reads and greps package.json, .mcp.json, and .project-standards.yaml files.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the processing templates.
Capability inventory: The skill has access to powerful tools including Bash, Write, and Edit, allowing for arbitrary command execution and file modification.
Sanitization: There is no evidence of sanitization or validation of the content read from external project files before it is processed by the agent.

configure-ux-testing