The Agent Skills Directory

[COMMAND_EXECUTION]: The helper script scripts/with_server.py executes arbitrary shell commands provided through the --server argument. It uses subprocess.Popen with shell=True to support complex command strings (e.g., using cd and &&), and also executes a final automation command using subprocess.run. While this is the primary purpose of the script for environment setup, it represents a direct command execution path.
[DYNAMIC_EXECUTION]: The skill's instructions encourage the agent to dynamically generate and execute Python Playwright scripts based on the state of the web application being tested. This involves the agent creating code at runtime and executing it via the with_server.py script.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from browser sessions, creating a vulnerability surface where malicious content in a web page could influence the agent's behavior.
Ingestion points: examples/console_logging.py captures all browser console messages; examples/element_discovery.py reads DOM content and element text.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the examples.
Capability inventory: The skill environment allows file writing (examples/console_logging.py) and arbitrary command execution (scripts/with_server.py).
Sanitization: There is no evidence of sanitization, filtering, or validation of the data captured from the browser before it is processed or logged.

webapp-testing