The Agent Skills Directory

[SAFE]: The skill consists of educational markdown files and code examples for software testing. No malicious scripts, hardcoded credentials, or exfiltration patterns were identified. A base64-encoded string in core/file-upload-download.md was verified as a benign 1x1 PNG image used for testing.
[PROMPT_INJECTION]: The skill's primary purpose is browser automation, which involves processing untrusted web content and creates a surface for indirect prompt injection. Ingestion points: Web pages navigated and read by the agent via playwright-cli (e.g., in playwright-cli/advanced-workflows.md). Boundary markers: There are no explicit markers or instructions to isolate web content from the agent's control logic. Capability inventory: Browser navigation, form interaction, JavaScript execution (run-code), and file operations. Sanitization: No sanitization of remote content is implemented. These capabilities are necessary for the skill's intended use and the risk is considered acceptable for the primary use case.
[EXTERNAL_DOWNLOADS]: Setup instructions refer to official registries and well-known repositories for testing tools. No suspicious or unverified download sources were found.

playwright-skill