ar-claim-builder

Construct evidence-grounded claim nodes and argument trees with hedge levels.

Updated Mar 22, 2026

Search Tags:#automation #workflow #research #evidence #debate #claim #argument

Ready to run

npx skills add https://github.com/ShinyGua/AutoArtsResearch --skill ar-claim-builder

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ar-claim-builder
Source: https://github.com/ShinyGua/AutoArtsResearch/tree/main/.claude/skills/ar-claim-builder
Command: npx skills add https://github.com/ShinyGua/AutoArtsResearch --skill ar-claim-builder

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Claim-building workflow that translates evidence into explicit claims and an auditable argument tree, enabling transparent reasoning and human review.

Core Features & Use Cases

Generate ClaimNodes grounded in evidence units (with identifiers and supporting/opposing evidence)
Assign hedge levels and detect overclaims to surface contested claims
Build a hierarchical argument tree mapping sub-claims to main thesis
Export outputs: claim files, argument_tree.json, and argument_summary.md for Gate 4 review
Supports human-in-the-loop decision gates during research workflows

Quick Start

Run the claim-builder on a prepared workspace to produce structured claims and the argument tree.

Frequently Asked Questions about ar-claim-builder

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

FAQPage Schema

How do I build an evidence-based argument tree from research data?▼

To build an evidence-based argument tree, this Skill constructs structured, evidence-grounded claims from provided evidence, mapping sub-claims to a main thesis while generating claim nodes and hedge levels for transparent reasoning.

What is a claim node and how does it support transparent reasoning?▼

A claim node is a structured unit grounded in specific evidence identifiers that maps supporting or opposing data to sub-claims. It supports transparent reasoning by making the evidentiary basis explicit and auditable for human review.

How do I detect overclaims and assign hedge levels during argument construction?▼

You detect overclaims and assign hedge levels by applying evidence-based discipline to the constructed claim nodes. This process surfaces contested claims and assigns appropriate hedge levels for subsequent human review.

Can I export an argument tree as JSON for automated analysis workflows?▼

Yes, you can export an argument tree as JSON. The workflow outputs an argument_tree.json file alongside claim files and an argument_summary.md, enabling automated analysis and integration into larger research workflows.

Does the argument builder support human-in-the-loop decision gates?▼

Yes, the argument builder supports human-in-the-loop decision gates. It generates debate needs and an argument summary specifically designed for Gate 4 review, ensuring transparent reasoning and human validation of contested claims.