The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill reads sensitive authentication credentials (email, password) from the project's internal memory (located at ~/.claude/projects/<project>/memory/reference_local_auth.md) and writes them into a local .env.test.local file. Although the skill includes instructions to update .gitignore to prevent these credentials from being committed to version control, the handling of plaintext secrets in local files is a sensitive operation.
[COMMAND_EXECUTION]: The skill dynamically generates TypeScript test files (.spec.ts) and executes them via npx playwright test. This pattern of script generation and subsequent execution creates a surface for indirect prompt injection if the source data files (such as app-map.md or playbooks) contain malicious content that influences the generated code.
[EXTERNAL_DOWNLOADS]: The skill performs automated installation of external dependencies, specifically the @playwright/test package and the Chromium browser via pnpm and npx. These resources originate from well-known and established service providers.

generate-tests