split-into-features

Decomposes an approved Epic into numbered Feature markdown files with acceptance criteria.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Breaking a large Epic into shippable slices is error-prone when done ad hoc: features end up oversized, overlapping, or missing acceptance criteria, and the Epic document loses track of its children. This Skill decomposes an approved Epic into consistently structured Feature files, each scoped to 1–2 weeks of solo work, and keeps the Epic's Features section linked and up to date. ## Core Features & Use Cases - Epic-to-Feature decomposition: Reads an Epic's Features outline and writes one FEAT-NN.M-<slug>.md file per feature under the Epic's folder, numbered sequentially. - Template-enforced structure: Uses FEATURE.template.md verbatim so every Feature has Goal, User value, Scope, testable Acceptance criteria, Dependencies, Data touched, and Risks. - Split governance: Applies an artefact-economy gate to avoid under- or over-decomposition, proposes merging or splitting oversized features, and asks for confirmation before writing. - Use Case: After approving EPIC-03 for a new onboarding flow, run the command to generate FEAT-03.1 through FEAT-03.4 with linked frontmatter and an updated Epic Features section, ready for story breakdown. ## Quick Start Ask the assistant to split EPIC-03 into features using /tandem:split-into-features and confirm the proposed file tree before anything is written.

Frequently Asked Questions about split-into-features

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

FAQPage Schema
How do I split an epic into features in a markdown-based project workflow?

Provide the path to the Epic file (or its EPIC-NN id) and invoke /tandem:split-into-features. The skill reads the Epic's Features outline, proposes a file tree for approval, then writes one FEAT-NN.M-<slug>.md file per feature under the Epic's folder.

How are feature files named and numbered?

Features are named FEAT-NN.M-<slug>.md, where NN is the Epic number and M increments sequentially within that Epic (.1, .2, .3). The skill scans the existing EPIC-NN folder to find the next free M before writing.

Does this skill also create stories for each feature?

No. Story creation is explicitly out of scope and handled by the next command, /tandem:split-into-stories, which decomposes each Feature into Stories and paired Testplans. This skill only writes Feature files and updates the Epic.

What happens if the epic only produces one feature?

The skill treats a single-feature result as an under-decomposition risk and stops to confirm the split count with the user before writing. It also proposes splitting any feature estimated larger than 1–2 weeks of work.

What frontmatter is set on each new feature file?

Each Feature gets status: not-started, created_at set to the current ISO 8601 timestamp with offset, empty strings for other timestamp fields, and an epic: field pointing back to the source EPIC-NN.