Hey Product Hunt,

I built LogicGuard because AI agents are getting good at drafting, but they still need a durable way to build and reuse the logic behind serious writing.

LogicGuard is an AI-facing logic-chain skill layer. It is not just a checker. Its subskills help an agent:

- preserve papers, books, reports, URLs, pasted text, and drafts into a source logic library;

- turn sources and drafts into reusable logic models: claims, evidence, warrants, assumptions, rebuttals, scope, and limitations;

- link project claims back to source nodes and anchored branches;

- audit whether the current logic chain is strong enough;

- synthesize new artifacts from existing models: paper outlines, technical reports, deck storylines, review responses, README positioning, and cautious rewrites.

The workflow I care about is:

find or collect material -> model the logic -> build a reusable model library -> select and reuse existing nodes -> synthesize a new paper, report, or storyline -> check where support is missing before writing too confidently.

So the core product is the skill system and model library that gives AI a memory of reasoning structure. The CLI, viewer, and source-library UI are support surfaces for preserving, inspecting, and moving those models.

It is especially useful for researchers, technical writers, and AI-agent users who work with many sources and need to produce papers, reports, briefs, or rebuttals without losing the logic chain.

I would love feedback on:

1. Is the model-library + synthesis story clear in the first minute?

2. Which output should the demo emphasize first: paper outline, technical report, deck storyline, review response, or AI-answer repair?

3. What kind of source-library UI would make reuse easiest?