refactor(commands): migrate execute sudolang program to slash command

[janhesters]

We might want to put this as /task instead.

I think even better would be to separate the slash commands. One for the /task part of the task creator, and one for the /execute command. WDYT?

---

Note

Moves the Task Creator guide from ai/rules/task-creator.mdc into ai/commands/execute.md, expanding it and deleting the original rules file.

• Commands:
  • ai/commands/execute.md: Replaces minimal execute doc with a full Task Creator guide, including Context Gathering, Requirements Analysis, Agent Orchestration, Task Planning, Execution Protocol, Task Plan Template, Completed Epic docs, Constraints, createTask pipeline, and Commands (/help, /task, /list).
• Rules:
  • Removes ai/rules/task-creator.mdc (content migrated).

^{Written by Cursor Bugbot for commit d6ac04c. This will update automatically on new commits. Configure here.}

[ericelliott]

@cursor 🙏 review

[cursor]

🔬 Code Review

I've completed a thorough review of the PR "refactor(commands): migrate execute sudolang program to slash command". Here are my findings:

Summary of Changes

• Moved: ai/rules/task-creator.mdc → ai/commands/execute.md
• Removed: YAML frontmatter metadata (description and alwaysApply fields)
• Changed: File extension from .mdc to .md

🚨 Critical Issues

1. Misalignment Between Command Structure and Implementation

The PR creates a structural inconsistency:

• please.mdc defines TWO commands (lines 36-37):

  • ✅ /task - plan and execute a task epic
  • ⚙️ /execute - execute a task epic (execution only)

• Current file structure:

  • ai/commands/task.md - stub that says "use the task creator to plan and execute"
  • ai/commands/execute.md - contains the full Task Creator implementation (both planning AND execution)

• Problem: execute.md contains the complete planning + execution workflow, but it should only contain execution logic per its command definition. The Commands section within execute.md (line 135) lists /task as a command, creating circular confusion.

📋 Recommendations

Option A: Merge into Single /task Command (Simpler)

1. Rename execute.md → task.md (overwrite the stub)
2. Remove /execute command from please.mdc
3. Keep the full Task Creator content in task.md
4. This aligns with the PR author's suggestion: "We might want to put this as /task instead"

Option B: Separate Planning and Execution (More Granular)

1. Split execute.md into two files:
   • task.md - contains planning workflow (createPlan() function)
   • execute.md - contains execution workflow (executePlan() function)
2. Update the Commands sections in each file accordingly
3. This implements the PR author's alternate suggestion: "separate the slash commands. One for the /task part...and one for the /execute command"

My Recommendation: Option A is preferable because:

• Simpler mental model for users
• The workflow already requires user approval between planning and execution
• Having separate commands adds unnecessary friction
• The createTask() function shows they're meant to work together: createPlan |> awaitAproval |> executePlan

✅ What's Working Well

1. Content preservation: All functionality from task-creator.mdc is preserved
2. No orphaned references: No files reference the old task-creator.mdc path
3. Clean migration: The content itself is unchanged and follows existing patterns
4. Constraints section: Properly maintained with clear guidelines

🎯 Action Items

Required:

1. Resolve command structure - Choose Option A or B above and implement it
2. Update file location - Ensure command files align with their definitions in please.mdc
3. Fix Commands section - Update line 135 in the file to reflect the correct command name

Nice to have:
4. Consider adding the YAML frontmatter back for consistency with other .md command files (if they use it)
5. Verify references to @agent-orchestrator.mdc (line 31) are still valid since this file has moved

Conclusion

The refactoring direction is sound, but the execution is incomplete. The command structure needs clarification before merging. I recommend Option A (merge into /task) as it simplifies the user experience while maintaining all functionality. The PR author's instinct that "We might want to put this as /task instead" is correct.

Would you like me to implement Option A or Option B to complete this refactoring?

---

 

_{Learn more about Cursor Agents}

[Copilot]

Pull Request Overview

This PR refactors the task execution workflow by migrating content from a rule file to a slash command. The content from task-creator.mdc has been moved to ai/commands/execute.md, transforming it from an always-applicable rule into an explicit command that users invoke.

Key Changes:

• Removed the entire task-creator.mdc rule file
• Replaced minimal command documentation in execute.md with full task creator implementation
• Changed from automatic rule application to explicit command invocation

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
ai/rules/task-creator.mdc	Completely removed the task creator rule file and all its content
ai/commands/execute.md	Replaced brief command description with full task creator implementation from the deleted rule file

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Missing space before the dash in command description. Should be '/list [(tasks|epics) = tasks] - list all tasks in the epic'.

[janhesters]

🧩 Refactor Request: Split Task Creator into Two Slash Commands

Hey @cursor 👋

Let’s implement Option B (separate planning and execution) from your review summary.

✅ Goal

Split the Sudolang Task Creator program into two slash commands:

1. /task → for planning (createPlan workflow)
2. /execute → for execution (executePlan workflow)

🧱 Required Changes

1. File Structure

• Move planning logic (createPlan, planTask, context gathering, etc.) into
  ai/commands/task.md
• Move execution logic (executePlan, validation, reporting, etc.) into
  ai/commands/execute.md

2. Command Declarations

• Update the Commands section at the bottom of each file:

  • task.md: only list /task
  • execute.md: only list /execute

3. Update please.mdc

• Ensure both commands are defined consistently:

  /task - plan a task or epic
  /execute - execute a previously planned task or epic

4. Code Adjustments

• Remove circular references between the two files (e.g., /execute listed inside task.md)

• Make sure createTask() correctly chains:

  createPlan |> awaitApproval |> executePlan
  

  where /task handles planning and /execute handles execution.

🧭 Acceptance Criteria

• /task runs full planning flow only (no execution)
• /execute runs execution flow only
• please.mdc reflects both commands cleanly
• No broken references or duplicated logic between files

[cursor]

Bug: Command Naming Conflicts with File Purpose

The execute.md file defines the /task command for creating tasks. This conflicts with the file's name and the PR's goal of separating task planning from task execution.

 

[ericelliott]

I'm good with separating /task and /execute - in fact, /execute feels like the wrong word. maybe /run [task] ?

[janhesters]

@ericelliott

I'm good with separating /task and /execute - in fact, /execute feels like the wrong word. maybe /run [task] ?

Done 👍