The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the mobile device's view hierarchy and local source code to determine navigation steps and element interactions.
Ingestion points: Data is retrieved via user-maestro-inspect_view_hierarchy (mobile UI state) and local files via search tools (grep and glob).
Boundary markers: No specific delimiters or safety instructions are defined to separate ingested UI text from the agent's internal logic.
Capability inventory: The skill can execute mobile automation commands (run_flow, run_flow_files), capture screenshots, and write temporary YAML files to the local file system.
Sanitization: The instructions do not specify any validation or sanitization of text elements extracted from the mobile application.
[DYNAMIC_EXECUTION]: The instructions direct the agent to dynamically generate Maestro flow files (.yml) containing UI interaction logic and then execute these generated files using the Maestro driver.
Evidence: The skill documentation suggests writing a "temp flow file in maestro/flows/_temp_explore.yml" and subsequently calling run_flow_files.

maestro-explore