The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface where malicious instructions could be embedded in the UI elements of an application being automated.
Ingestion points: The agent ingests data from the desktop accessibility tree (UIA/AT-SPI2) via XPath queries and attribute retrieval as described in references/locators-and-queries.md.
Boundary markers: There are no instructions or delimiters provided to help the agent distinguish between UI metadata and untrusted text content within the application UI.
Capability inventory: The skill provides full interaction capabilities, including Keyboard Type and Pointer Click (documented in references/keywords-reference.md), which could be misused if the agent obeys instructions found in the UI.
Sanitization: No validation or filtering mechanisms are documented for data retrieved from UI nodes before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill directs the user to install the PlatynUI library and its associated CLI tools from the public Python package registry.
Evidence: SKILL.md and references/cli-and-inspector.md provide pip and uv commands to install robotframework-PlatynUI (pinning version 0.12.0.dev330), platynui-cli, and platynui-inspector.
[COMMAND_EXECUTION]: The skill utilizes external libraries to execute local system processes.
Evidence: The example in assets/examples/calculator.robot uses the Robot Framework Process library to execute calc.exe as part of the test setup.

rf-platynui