# To Issues

> Break a PRD or plan into independently grabbable tracer-bullet issues.

- Skill: `youssefguba/to-issues` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add youssefguba/to-issues`
- Raw SKILL.md: https://api.skillmd.com/api/skills/youssefguba/to-issues/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: Youssefguba (https://skillmd.com/u/youssefguba)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/youssefguba/to-issues

---


# To Issues

Break a PRD, plan, or conversation into **tracer-bullet** issues. Each issue is
a narrow but complete path through the relevant integration layers and is
demoable or verifiable on its own.

## Process

### 1. Gather context

Read the full source PRD, issue, comments, and relevant codebase guidance. Use
the project's domain vocabulary and respect architecture decisions.

### 2. Draft vertical slices

For every proposed issue, capture:

- **Title**: short and outcome-focused.
- **Blocked by**: only issues that genuinely gate the work.
- **User stories covered**: source PRD story numbers, when available.
- **What it delivers**: observable end-to-end behavior.

Prefer many thin vertical slices over a few horizontal layer tasks. Size each
slice to fit a fresh agent context. When a wide mechanical refactor cannot land
green as a vertical slice, use an expand, migrate, contract sequence.

### 3. Quiz the user

Present the numbered breakdown and ask whether the granularity and dependency
edges are correct. Iterate until approved.

### 4. Publish in dependency order

Create blockers first so later issues can reference real identifiers. Apply the
`ready-for-agent` triage label only to the current unblocked frontier unless
instructed otherwise. Promote successors to that label as their blockers
complete. Do not close or modify a parent PRD issue.

Use this issue body:

```markdown
## Parent

Reference the source PRD issue when one exists.

## What to build

Describe the end-to-end behavior, not a layer-by-layer implementation list.

## Acceptance criteria

- [ ] Observable criterion 1
- [ ] Observable criterion 2

## Blocked by

- Reference each blocking issue, or state that work can start immediately.
```

The implementation frontier contains every issue whose blockers are complete.