Remove "I will" phrasing from plan templates to make co-editing easier

[Copilot]

Problem

Plans created in Magentic-UI were unnecessarily verbose and difficult to co-edit due to the repetitive "I will" phrasing at the beginning of each step.

From the screenshot in the issue, we can see that steps all started with phrases like "I will search for..." and "I will access the menu information..." which added unnecessary wordiness.

Solution

This PR updates the plan templates to use more concise, direct language:

1. Modified the prompt instructions in _prompts.py to remove the requirement for first-person phrasing

2. Updated all example steps in the templates to use a direct style that:

   • Starts with a brief action verb matching the title
   • Avoids redundant "I will" statements
   • Preserves all the important information in a more concise format

3. Changed the JSON schema examples to reflect the new style

Before vs After

Before:

- title: "Locate the menu of the first restaurant"
- details: "I will locate the menu of the first restaurant. \n This involves searching for top-rated restaurants in the 98052 area, selecting one with good reviews and an accessible menu, then extracting and formatting the menu information."

After:

- title: "Locate the menu of the first restaurant" 
- details: "Locate the menu of the first restaurant. \n Search for top-rated restaurants in the 98052 area, select one with good reviews and an accessible menu, then extract and format the menu information."

The changes maintain all necessary information while making the plans more readable and easier to edit.

Fixes #37.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[gagb]

Changes look good to me and tested the PR and it works. It basically changed the prompts.

[gagb]

@husseinmozannar and I discussed and he will revisit this in a week or so.

[matheusmaldaner]

Properly addresses the Issue #37