split-into-stories

Decompose a Feature into Stories with paired Testplans in one response.

Updated May 25, 2026
One-click install
npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill split-into-stories-data-ai-xyz
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: split-into-stories
Source: https://github.com/DATA-AI-XYZ/Tandem/tree/main/skills/split-into-stories
Command: npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill split-into-stories-data-ai-xyz

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Breaking an approved Feature into work-sized Stories often produces orphan Stories with no test coverage, or Testplans written later that drift from the acceptance criteria. This Skill enforces a mandatory Story-to-Testplan pairing rule: both artefacts are drafted, validated, and written to disk in the same response, or nothing is written at all. ## Core Features & Use Cases - Paired Story and Testplan generation: Every Story file (STORY-NN.M.PP) ships with a lockstep-numbered Testplan (TESTPLAN-NN.M.PP) where each acceptance criterion maps to at least one test case with a runnable command. - Structural abort on untestable criteria: If an acceptance criterion is not machine-testable, the Skill stops, flags the offending AC, and writes nothing rather than leaving an orphan Story. - Template and path-map driven output: Uses STORY.template.md and TESTPLAN.template.md verbatim, resolves folders via pm-paths.js, fills frontmatter (status, epic, feature, type_of_work), and runs a usage-estimate script per story. - Use Case: After approving FEAT-03.2, ask the Skill to decompose it; it reads the Feature's acceptance criteria, proposes the fewest Stories that carry the testable claims, shows the file tree for approval, then writes Stories, Testplans, and updates the Feature's Stories section. ## Quick Start Ask the assistant to split the approved Feature file FEAT-03.2 into Stories with paired Testplans using the tandem split-into-stories command.

Frequently Asked Questions about split-into-stories

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I split a feature into user stories with test plans?

Provide the path to the approved Feature file and invoke the split-into-stories command. The Skill reads the Feature's acceptance criteria, drafts each Story and its paired Testplan in memory, shows the proposed file tree for approval, then writes both artefacts in the same response.

What happens if an acceptance criterion is not machine-testable?

The Skill aborts and writes nothing. It flags the specific acceptance criterion, explains why it is not testable, and asks you to rewrite it as a falsifiable statement or accept dropping it, rather than leaving an orphan Story without a Testplan.

How are Story and Testplan files numbered and organized?

Stories and Testplans use lockstep numbering: STORY-NN.M.PP pairs with TESTPLAN-NN.M.PP for the same PP, zero-padded to two digits. Files live under EPIC-NN/FEAT-NN.M/ subfolders resolved through the pm-paths.js path map rather than hardcoded locations.

Can a story be marked ready immediately after splitting?

No. Stories are created with status not-started and ai_review pending. Promotion to ready requires passing the Definition of Ready gate through the separate refine-backlog step, though an obviously clean story can go straight to execute-story.

When should a story be split further during decomposition?

Split when a story estimates to XL, contains more than five acceptance criteria, or when a single criterion implies more than a large unit of work. Conversely, criteria without their own falsifiable claim fold into a sibling Story to keep artefact count minimal.