The Agent Skills Directory

[SAFE]: The skill is a legitimate testing tool for UI functionality. It utilizes standard local development environment setups (localhost) and browser automation tools (Playwright) for verification. The instructions and test data are consistent with the stated purpose of software quality assurance.
[INDIRECT_PROMPT_INJECTION]: The skill defines a file upload test case that requires the agent to process and summarize the contents of various files (HTML, XML, JSON, etc.). This pattern constitutes a structural surface for indirect prompt injection, as the agent is asked to interpret data that could theoretically contain instructions. In this specific skill, the provided test files are benign samples used solely for validation.
Ingestion points: Multiple file formats in the test-files/ directory (CSV, HTML, JSON, MD, TXT, XML) are used as input for the file_upload validation task.
Boundary markers: The instructions do not implement specific delimiters or 'ignore instructions' warnings to isolate the file content from the agent's task context during summarization.
Capability inventory: The skill utilizes Playwright MCP tools and browser navigation capabilities to interact with the UI.
Sanitization: No sanitization or content filtering is specified for the data contained within the uploaded test files.

validating-ui-features