▶
Executionexec_in
Trigger
Automation/LLM/Planning
Uses LLM to suggest the most appropriate next action given current screen and goal
Scores range from 0 to 10. Higher values mean more impact, exposure, or operational weight.
Trigger
Vision-capable LLM model
Base64-encoded current screenshot
Ultimate goal we're trying to achieve
JSON array of actions already taken
Result/outcome of the last action
Continue
Goal appears to be achieved
Next step suggestion
Type of suggested action
Target description