textum

Orchestrate a four-phase PRD to Story workflow with gate checks.

40|3|Updated Jan 1, 2026
One-click install
npx skills add https://github.com/snakeying/Textum --skill textum
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: textum
Source: https://github.com/snakeying/Textum/tree/main/.codex/skills/textum
Command: npx skills add https://github.com/snakeying/Textum --skill textum

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Textum provides a structured four-phase workflow to reduce AI drift by gating PRD planning, scaffolding, splitting, and story execution behind deterministic stages.

Core Features & Use Cases

  • PRD Plan: transform fuzzy ideas into a clear, versioned requirements document.
  • Scaffold: render a global context and architectural decisions that stay consistent across stages.
  • Split: break requirements into manageable stories with clear ownership and dependencies.
  • Story Exec: execute one story at a time with gate checks and validations to ensure quality.

Quick Start

Describe your idea in plain language to start the four-phase workflow with built-in gate checks.

Frequently Asked Questions about textum

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

FAQPage Schema
How do I prevent unpredictable AI generations when planning a software feature?

To prevent unpredictable AI generations, use a structured PRD workflow with deterministic gate checks that enforce strict input/output contracts across planning, scaffolding, splitting, and story execution stages.

What is the best way to break down a large PRD into manageable development stories?

The best way to break down a PRD is to use a split phase that decomposes requirements into manageable stories with clear ownership and dependencies, ensuring consistent global context across all tasks.

How do I maintain consistent architectural decisions across multiple AI-generated stories?

You maintain consistent architectural decisions by using a scaffold phase that renders global context and architecture choices, keeping them consistent across subsequent splitting and story execution stages.

Does a gated PRD workflow support reproducible AI task execution?

A gated PRD workflow supports reproducible AI task execution by enforcing deterministic stages with validations and producing reusable artifacts, ensuring each story generates consistent results.

When should I avoid using an end-to-end PRD scaffolding workflow?

You should avoid an end-to-end PRD scaffolding workflow for simple, single-step tasks that do not require architectural decisions, task decomposition, or strict deterministic context validation.