The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to perform application testing which includes running CLI commands directly and using platform-bundled tools (browser-use@openai-bundled and computer-use@openai-bundled) for UI interaction. These actions are within the expected scope of an automated testing skill.
[INDIRECT_PROMPT_INJECTION]: The skill operates by reading test scenarios from local files located in '.turbo/test-plans/'. While these files define the steps for the agent to follow, this is the intended design for providing test instructions. Mandatory Evidence Chain: (1) Ingestion points: Test plan files in '.turbo/test-plans/'; (2) Boundary markers: The agent is instructed to 'Perform the exact steps' defined in the scenario, though no specific isolation markers for the file content are mentioned; (3) Capability inventory: Execution of CLI commands and use of browser/system control plugins; (4) Sanitization: The skill does not implement specific sanitization of the scenario steps, relying on the agent's interpretation of the markdown instructions.

exploratory-test