Triage New Issues

When a new issue is opened, perform the following triage steps.

Steps

1. Read the issue title, body, and any attached context carefully

2. Classify the type based on content:

   • Describes broken behavior → bug
   • Requests new capability → feature
   • Asks a question → question
   • Relates to docs → documentation
   • Relates to infrastructure → type/infra

3. Classify the area based on files or systems mentioned:

   • Frontend / UI → frontend
   • API / backend → backend, api
   • CI/CD → devops
   • Canvas/drawing → canvas

4. Assign priority based on impact described:

   • Data loss, security, or complete breakage → priority-critical
   • Significant user impact, blocked workflow → priority-high
   • Degraded experience, workaround exists → priority-medium
   • Minor inconvenience, nice to have → priority-low

5. Apply labels from the above classifications

6. Search for related issues using keywords from the title

   • If related issues found, add a comment: "Related to #{N} — [title]"

7. Ask clarifying questions if the issue is missing:

   • Steps to reproduce (for bugs)
   • Expected vs actual behavior (for bugs)
   • Use case or motivation (for features)
   • Environment details if relevant

   Post as a single comment with all questions at once — not multiple comments.

8. Assign to milestone if the issue clearly belongs to an active milestone

Comment Template for Clarification

Thank you for opening this issue!

To help us address this effectively, could you provide:

- [ ] Steps to reproduce the issue
- [ ] What you expected to happen
- [ ] What actually happened
- [ ] Your environment (OS, browser, version)

The more detail you can share, the faster we can help.

*— The Doorman, Agenthood*

Notes

• Never close an issue during triage — only label and comment
• If the issue is clearly a duplicate, add duplicate label and link to the original
• This workflow runs with read-only permissions; labels and comments are safe outputs