The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection.
Ingestion points: The skill navigates to and processes external websites and DOM structures provided by users or found via navigation (SKILL.md, Visual Inspection Phase).
Boundary markers: There are no specific instructions or delimiters provided to the agent to distinguish between its own system instructions and potentially malicious instructions embedded in the processed web content.
Capability inventory: The agent has extensive capabilities including reading and writing files in the workspace and searching the codebase (SKILL.md, Required Capabilities).
Sanitization: No sanitization or validation of the retrieved web content is implemented before it is analyzed by the agent.- [EXTERNAL_DOWNLOADS]: Recommends the installation and use of the official Playwright MCP server.
Evidence: The configuration example uses npx -y @playwright/mcp@latest (SKILL.md, Reference Implementation).
Context: This is a well-known service provided by Microsoft and is considered a safe dependency for the skill's purpose.- [COMMAND_EXECUTION]: The skill workflow involves executing browser automation and code search tools.
Evidence: Uses tools like browser_navigate, grep_search, and semantic_search to interact with websites and the local codebase.

web-design-reviewer