skills.rest
curdx
curdxCOMMUNITY
Explore all skills by this author →
GitHub

new

Create a curdx-flow spec directory with validated state and research output.

4|Updated Apr 26, 2026
Search Tags:#multi-directory#research phase#requirements planning#spec creation#Claude Code plugin#state initialization
One-click install
npx skills add https://github.com/curdx/curdx-flow --skill new-curdx
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: new
Source: https://github.com/curdx/curdx-flow/tree/main/plugins/curdx-flow/skills/new
Command: npx skills add https://github.com/curdx/curdx-flow --skill new-curdx

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns a single spec name and goal into a structured, durable spec folder that Claude Code can continue across restarts while enforcing validation and capturing the original goal.

Core Features & Use Cases

  • Creates a new spec directory with consistent state: Generates the spec folder, .curdx-state.json, .progress.md, and updates .current-spec and .gitignore entries.
  • Captures and validates inputs before execution: Enforces kebab-case spec names, supports optional --skip-research, and validates --specs-dir against configured roots.
  • Starts research with a dedicated agent and stops for review: Runs research-analyst to produce research.md, then stops to let the user verify findings before requirements begin.

Quick Start

Run /curdx-flow:new user-auth Add OAuth2 login and then review the generated research.md before continuing.

Frequently Asked Questions about new

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

FAQPage Schema
How do I create a spec directory with resumable state for Claude Code?▼

To create a spec directory with resumable state, provide a spec name and goal to initialize a structured folder containing .curdx-state.json and .progress.md, allowing Claude Code to safely resume delivery across restarts.

How do I start the research phase for requirements planning and skip it if needed?▼

To start the research phase for requirements planning, the Skill triggers an agent-driven research-analyst to output research.md. You can skip this phase by passing the optional --skip-research flag during spec creation.

Does spec creation enforce kebab-case names and validate the specs directory?▼

Yes, spec creation enforces kebab-case name validation for the provided spec name and validates the --specs-dir argument against configured roots to ensure multi-directory support and safe state initialization.

What is the workflow after the research phase generates research.md?▼

After the research phase generates research.md, the workflow halts execution. This stop mechanism lets you review and verify the research findings before beginning the requirements planning phase.

Can I use this for multi-directory spec discovery workflows?▼

Yes, you can use this for multi-directory spec discovery and planning workflows. It validates the target specs directory against configured roots and safely initializes state for controlled generation.

Related Skills

Explore More Product & Management Skills→
biwakonbubiwakonbu

skill-spec

Validate SKILL.md frontmatter and structure for Claude Code plugin Skill Units.

Community
Basic
curdxcurdx
4

design

Generate design.md from requirements.md with two-stage spec compliance review.

Community
Advanced
NetzSichtNetzSicht

kategorie-research-template

Create a structured kategorie-research.md brief for a new product category.

Official
Intermediate
jhillbhtjhillbht

skill-creator-enhanced

Create and package Claude skills into distributable .skill files.

Community
Advanced