The Agent Skills Directory

[COMMAND_EXECUTION]: The helper script scripts/with_server.py uses subprocess.Popen with shell=True to execute server startup commands and subprocess.run to execute the final automation command. This is a functional requirement to support shell features like command chaining (e.g., cd backend && python server.py) and environment setup during local development testing.
[PROMPT_INJECTION]: The SKILL.md instructions explicitly direct the agent not to read the source code of the helper scripts unless necessary, citing context window limitations. While this is a common optimization for LLM agents, it effectively asks the agent to treat the provided scripts as black boxes, reducing transparency of executed logic.

webapp-testing