stories

name	stories
description	Use this skill when the user asks to "create stories from the typescript refactor plan", "emit stories for ts migration", "scrum-master stories from migration plan", "ts refactor stories", "hand off the typescript plan to scrum master", or has an existing migration plan (or migration-slices report) on disk and wants per-story scrum-master-format markdown files generated. Reads a plan file (or the most recent blackboard plan) and emits one story file per slice in the exact scrum-master schema, with dependency rewrites, AC validation, and Mermaid DAG. Read-mostly -- writes stories to the scrum-master board path. For a full plan from scratch, use /typescript-refactor.
argument-hint	[plan-file-path | "latest" | --board=<path>]
allowed-tools	Bash, Read, Write, Edit, Grep, Glob, TodoWrite, mcp__plugin_serena_serena__activate_project

TypeScript Refactor -- Stories Handoff

You are running the stories handoff. You take an existing migration plan (or the slicer's per-slice catalog) and write one scrum-master-format story file per slice.

Your job:

1. Locate the plan
2. Resolve the board path
3. Validate the plan's slice catalog
4. Write one story file per slice
5. Write the Mermaid DAG
6. Summarize

Step 1: Locate the plan

Argument	Action
empty or latest	Find newest .claude/typescript-refactor/*/plan-final.md
<path>	Use the path verbatim
--board=<path>	Use the path AND override the board target

If no plan exists: tell the user to run /typescript-refactor first. Stop.

Step 2: Resolve the board path

1. Inline --board=<path> argument
2. .claude/scrum-master.local.md board_path: field
3. .claude/typescript-refactor.local.md scrum_master_board_path: field
4. Auto-detect: glob **/Backlog/current/
5. Default: <PROJECT_ROOT>/Backlog/current/

Step 3: Validate the slice catalog

Parse the plan's slice catalog. Verify required fields exist. FAIL on missing title, scope.include, acceptance with prose-only AC, or missing rollback.

Step 4: Scan existing board for ID collisions

On collision, append A suffix and rewrite all dependency references.

Step 5: Emit one story file per slice

Read the canonical schema: ${CLAUDE_PLUGIN_ROOT}/skills/typescript-refactor/references/story-schema.md. Instantiate it for each slice.

Step 6: Write Mermaid DAG

Generate <BOARD_PATH>/../board-dag.md with color-coded nodes per epic.

Step 7: Summary

Do not print the summary until all three verifications pass:

1. ls <BOARD_PATH>/TSR-*.md | wc -l equals the slice count validated in Step 3
2. Every id in every emitted story's blocks/blocked-by list has a matching story file on the board
3. The DAG file contains a ```mermaid fence with one node per emitted story

Stories handoff complete

Source plan:    <path>
Board path:     <abs>
Stories written: <N>

By epic:
  Foundation (F):   <n>
  Boundaries (B):   <n>
  Renames (R):      <n>
  Framework (FW):   <n>
  Cleanup (C):      <n>

Next:
  /scrum-master status     # see populated board
  /scrum-master deps       # render the DAG

Anti-patterns

• Do NOT modify the source plan
• Do NOT skip AC validation
• Do NOT emit stories with empty scope.include
• Do NOT emit stories without rollback
• Do NOT silently drop dependencies when ID collisions force renames