Skip to content

Comments

agent: update make-task-issue to use prompt-based input instead of editor#20

Merged
p-iknow merged 1 commit intomainfrom
i19-agent/prompt-based-input-for-make-task-issue
Oct 25, 2025
Merged

agent: update make-task-issue to use prompt-based input instead of editor#20
p-iknow merged 1 commit intomainfrom
i19-agent/prompt-based-input-for-make-task-issue

Conversation

@p-iknow
Copy link
Contributor

@p-iknow p-iknow commented Oct 25, 2025

Requirement

Resolves #19

Update the /make-task-issue command to use a prompt-based input approach instead of editor-based input. The current editor-based approach doesn't work properly in some environments and requires complex editor configuration, degrading user experience.

Implementation

Updated .claude/commands/make-task-issue.md to replace editor-based workflow with prompt-based approach:

  • Section 2 restructured: Changed from "Editor-based" to "Prompt-based" input collection
  • Type-specific prompts defined: Created sequential prompts for Context, Requirement, and optional sections tailored to each issue type (feat, doc, test, etc.)
  • Validation rules added: Required sections (Context, Requirement) validate for non-empty input and re-prompt if validation fails
  • Editor dependencies removed:
    • Removed Write tool from allowed-tools frontmatter (no longer needed for temp files)
    • Removed all ${EDITOR:-vi} implementation steps
    • Removed temporary file creation logic
  • Examples updated: Changed from showing editor content to Q&A prompt format
  • Error handling updated: Changed from "re-open editor" to "re-prompt for section"
  • Documentation updated: Implementation Notes now reflect prompt-based approach

The new approach follows the proven pattern from /make-command, providing type-specific prompts for all 9 issue types while maintaining the AI Enhancement functionality.

Test

Manual testing recommended:

  • Test /make-task-issue feat - verify all prompts appear correctly
  • Test required section validation - verify re-prompting when empty
  • Test optional section skipping - verify Enter to skip works
  • Verify AI enhancement flow still works properly
  • Test with different issue types (doc, test, agent) to verify type-specific prompts

Human Check

🤖 Generated with Claude Code

  • I reviewed it

…itor

Requirement:
The /make-task-issue command currently uses an editor-based approach which doesn't work properly in some environments and requires complex editor configuration. Update to use a prompt-based input approach for better reliability and user experience.

Implementation:
- Changed Section 2 from "Editor-based" to "Prompt-based" collection
- Defined type-specific sequential prompts for Context, Requirement, and optional sections
- Added validation rules for required sections with re-prompting behavior
- Removed Write tool from allowed-tools (no longer needed for temp files)
- Removed all editor-related implementation steps (Write tool, ${EDITOR:-vi} calls)
- Updated examples to show Q&A format instead of editor content
- Updated error handling to use re-prompting instead of re-opening editor
- Updated Implementation Notes to reflect prompt-based approach

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
@github-actions github-actions bot added the agent Agent rules and commands label Oct 25, 2025
@p-iknow p-iknow marked this pull request as ready for review October 25, 2025 14:37
@p-iknow p-iknow merged commit 336c3dc into main Oct 25, 2025
4 checks passed
@p-iknow p-iknow deleted the i19-agent/prompt-based-input-for-make-task-issue branch October 25, 2025 14:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

agent Agent rules and commands

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

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

1 participant