construct-domain-init

Configures research domains by populating domains.yaml and search-seeds.json through an interactive interview.

Updated Apr 21, 2026
One-click install
npx skills add https://github.com/mabstruct/construct --skill construct-domain-init-mabstruct
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: construct-domain-init
Source: https://github.com/mabstruct/construct/tree/main/CONSTRUCT-CLAUDE-impl/claude/skills/construct-domain-init
Command: npx skills add https://github.com/mabstruct/construct --skill construct-domain-init-mabstruct

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Setting up a new research domain in a CONSTRUCT workspace requires manually editing YAML and JSON configuration files with the correct schema, which is error-prone and tedious. This Skill guides users through a structured interview and writes valid domain configuration automatically. ## Core Features & Use Cases - Interactive Scope Interview: Asks adaptive questions about domain topics, key authors, source priorities, and cross-domain overlaps, skipping anything the user already answered. - Config Generation: Converts the domain name to a unique kebab-case ID and writes validated entries to domains.yaml and weighted search clusters to search-seeds.json. - Workspace Bootstrapping: Creates the domain's digests/ directory, logs an event to log/events.jsonl, and can chain directly from workspace-init. - Use Case: A researcher says "add a domain for climate adaptation policy" and the Skill interviews them, then produces a ready-to-use domain configuration with search seed clusters so a research cycle can start immediately. ## Quick Start Say "add a domain" or "initialize a new research domain" and answer the interview questions to generate your domain configuration.

Frequently Asked Questions about construct-domain-init

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

FAQPage Schema
How do I add a new research domain to my CONSTRUCT workspace?▼

Say "add a domain" or "initialize domain" to trigger an interactive interview covering topics, key authors, and source priorities. The Skill then writes the configuration to domains.yaml and search-seeds.json in your domain workspace subdirectory.

What files does domain initialization create or modify?▼

It adds an entry to domains.yaml with description, content categories, and source priorities, plus 2-5 weighted search clusters in search-seeds.json. It also creates a digests/{domain-id}/ directory and appends an event to log/events.jsonl.

Can I initialize a domain without knowing key papers or authors?▼

Yes, the interview treats known papers, authors, and institutions as optional input for users starting cold. The Skill still creates a core topic cluster from your subject description so research cycles can begin immediately.

How are search seed clusters weighted during domain setup?▼

The core topic cluster receives the highest weight (0.8), while author and institution clusters receive lower weights (0.5). Between 2 and 5 clusters are created depending on interview depth.

What validation happens after domain configuration is written?▼

The Skill checks that the domain ID is unique, at least 2 content categories exist, at least 1 search cluster was created, and both domains.yaml and search-seeds.json remain syntactically valid after the updates.