Summary
Allow users to modify or revise any prompt they’ve queued before it executes, enabling on-the-fly adjustments based on evolving context or new insights.
Problem Statement
- Current Behavior: Users can queue multiple prompts in a single conversation with Cursor Agent, but once a prompt is queued, its content is locked in.
- Pain Point: As the Agent processes earlier prompts and generates partial outputs, users often realize they need to tweak or clarify subsequent queued prompts. Without editability, they must cancel and re-queue prompts or wait for unwanted results, disrupting workflow and context.
Proposal
Introduce an “Edit” option for queued prompts in the chat UI. Before a prompt is sent to the Agent, users can click an inline edit icon (✎) to update the text, then save or cancel their changes.
Key UX Changes
- Inline Edit Icon: Displayed next to each pending prompt in the queue.
- Edit Mode: Clicking the icon turns the prompt bubble into an editable text area with “Save” and “Cancel” actions.
- Change Tracking: Optionally highlight edited text to indicate revisions.
- Keyboard Shortcuts: Support ⌘/Ctrl+Enter to save edits and Esc to cancel.
User Stories
- As a researcher, I want to refine the wording of a queued query after seeing preliminary results, so I can steer the Agent toward more precise answers.
- As a developer, I want to correct an overlooked parameter in a later prompt without restarting the entire session, saving time and preserving context.
- As a data engineer, I want to clarify ambiguous field names in my queued requests after reviewing partial outputs, ensuring accurate data workflows.
Benefits
- Improved Flexibility: Users can adapt to new information in real time without breaking the conversational flow.
- Time Savings: Eliminates the need to delete and re-enter prompts, reducing friction and cognitive load.
- Enhanced Precision: Enables iterative refinements, leading to higher-quality outputs from Cursor Agent.