The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches standard testing libraries and tools from well-known package registries, including Playwright, axe-core, and jest-axe. This is normal behavior for a testing-oriented development skill.
[COMMAND_EXECUTION]: Recommends standard Playwright CLI commands for running tests, initializing AI agents, and setting up MCP servers. These commands are consistent with official documentation and the primary purpose of the skill.
[PROMPT_INJECTION]: Analyzed for indirect prompt injection surfaces as the skill is designed to interact with and test external URLs; however, the instructions emphasize best practices such as mocking external APIs to maintain deterministic and safe test environments.
[SAFE]: Utilizes dynamic context injection (!command syntax) in utility scripts to perform benign local reconnaissance, such as counting existing page objects and detecting the project's test directory. These operations are non-destructive and limited to local metadata gathering.

testing-e2e