The Agent Skills Directory

Unverifiable Dependencies & Remote Code Execution (LOW): The skill requires pip install playwright and playwright install chromium. While Playwright is a trusted, standard industry tool, it involves downloading and executing external binaries.
Indirect Prompt Injection (LOW): The skill processes untrusted data from external websites, which could contain malicious instructions designed to manipulate the AI agent.
Ingestion points: scripts/playwright_runner.py ingests data from arbitrary URLs via page.goto() and extracts elements like page.title().
Boundary markers: Absent. The script returns raw content from the web page without delimiters or instructions to the agent to ignore embedded commands.
Capability inventory: The skill is granted Bash, Write, and Edit permissions, providing a high-impact target if an injection succeeds.
Sanitization: None. The script extracts and returns page metadata and text directly to the agent's context.
Command Execution (SAFE): The skill uses Bash to run its internal Python scripts. The logic in scripts/playwright_runner.py is focused on browser automation and does not exhibit malicious command construction or privilege escalation patterns.

webapp-testing