Skip to content

agent: update make-task-issue to use prompt-based input instead of editor #19

Description

@p-iknow

Context

The /make-task-issue command currently uses an editor-based input approach, which doesn't work properly in some environments or requires complex editor configuration. This degrades the user experience.

Codebase analysis findings:

  • .claude/commands/make-task-issue.md:49-125 - Currently uses ${EDITOR:-vi} to create temporary file and open editor
  • .claude/commands/make-command.md:47-100 - Already successfully uses prompt-based approach (Q1-Q4 interactive questions)
  • Error handling section mentions "If editor cannot be opened, fall back to multi-line prompts" but not implemented

Requirement

Update the /make-task-issue command to use a prompt-based input approach, allowing users to input each section through interactive prompts.

Key requirements:

  • Remove editor dependency
  • Provide type-specific sequential questions
  • Validate required sections (Context, Requirement)
  • Allow optional sections to be skipped
  • Maintain AI Enhancement functionality

Solution

Implement following /make-command's prompt-based pattern:

1. Section-by-section sequential input:

  • Context (required): "Describe the background: Why is this needed? What is the current situation?"
  • Requirement (required): "Describe what needs to be done and the expected outcome:"
  • Solution (optional): "If you have ideas about how to solve this, describe the approach: (Enter to skip)"
  • Test Plan/Verification/Content Outline (optional, type-specific): "How should this be tested or validated? (Enter to skip)"

2. Type-specific question mapping:

  • feat, config, agent: Context → Requirement → Solution → Test Plan
  • doc: Context → Requirement → Content Outline → Verification
  • test: Context → Requirement → Test Scope → Verification
  • perf, refactor: Context → Requirement → Solution → Verification
  • fix: Context → Requirement → Solution → Test Plan

3. Implementation changes:

  • Replace entire Section 2 "Collect User Input (Editor-based)"
  • Remove Write tool logic for temporary file creation
  • Remove Bash(${EDITOR:-vi}) calls
  • Collect each section via interactive prompts
  • Validate required fields (re-prompt if empty or whitespace only)

4. Maintain AI Enhancement flow:

  • After user input collection completes
  • Gather related context from codebase
  • Present AI suggestions to user for approval

Test Plan

  1. Test issue creation with various types:

    • /make-task-issue feat - Fill all sections
    • /make-task-issue doc - Fill required only, skip optional sections
    • /make-task-issue agent - Test AI enhancement approval/rejection
  2. Verification items:

    • Confirm re-prompt when required sections are empty
    • Confirm optional sections can be skipped
    • Verify correct questions shown per type
    • Verify AI enhancement flow works properly
    • Verify final issue structure and content

Reference

  • .claude/commands/make-task-issue.md - Current implementation
  • .claude/commands/make-command.md - Reference prompt-based pattern

Metadata

Metadata

Assignees

Labels

agentAgent rules and commands

Type

Fields

No fields configured for Task.

Projects

Status
Done

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions