theory-ingest

Extracts assumptions, propositions, and testable implications from theory papers into a structured research wiki.

Updated Jul 15, 2026
One-click install
npx skills add https://github.com/1549171506/shizhengmubangai1 --skill theory-ingest-1549171506
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: theory-ingest
Source: https://github.com/1549171506/shizhengmubangai1/tree/main/Desktop/123/empiricalwiki/i18n/en/skills/theory-ingest
Command: npx skills add https://github.com/1549171506/shizhengmubangai1 --skill theory-ingest-1549171506

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pymupdf.

What problem does it solve? Turning a dense theory-modeling paper into reusable research assets is slow and error-prone: formal statements get paraphrased, assumptions get lost, and links to empirical work are never recorded. This Skill converts one theory paper (PDF or TeX) into structured wiki pages with verbatim formal statements and a shared graph connecting theory to the empirical layer. ## Core Features & Use Cases - Six-slot extraction: Pulls environment/primitives, solution concept, results, proofs, comparative statics, and testable implications along a fixed theory-modeling skeleton, writing "not reported" instead of guessing. - Structured wiki output: Creates or updates paper cards, assumptions, propositions, hypotheses, mechanisms, and foundations pages with verbatim formal statements and explicit deduplication. - Knowledge graph edges: Records assumes/proves/formalizes_mechanism/predicts edges with confidence levels, bridging theory to empirical papers on one shared graph. - Use Case: Given a TeX source of a game-theory paper on managerial myopia, generate a paper card plus proposition pages with verbatim statements, link its testable implications to hypothesis pages, and connect the formalized mechanism to existing empirical mechanisms. ## Quick Start Run the theory-ingest skill on the local file paper.tex with the topic "managerial myopia and patient capital" to generate the wiki pages and graph edges.

Frequently Asked Questions about theory-ingest

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

FAQPage Schema
How do I extract propositions and assumptions from a theory paper?

Run the theory-ingest workflow on the paper's PDF or TeX source. It extracts content along six fixed slots and writes each core assumption to assumptions pages and each proposition to propositions pages with the formal statement quoted verbatim and proof technique recorded.

Should I use PDF or TeX source for ingesting theory papers?

TeX source is strongly preferred because theory papers are formula-dense and equations survive extraction. With only a PDF, the workflow reads the first page for the title using PyMuPDF, then prepares the source through a preparation script.

What is the difference between theory-ingest and empirical-ingest?

Theory-ingest handles formal models with assumptions, propositions, and proofs, while empirical-ingest handles variables, datasets, identification, and robustness. Theory papers have no empirical blocks, and empirical papers write the reverse test edges back to theory hypotheses.

What happens when a proof technique or result is not stated in the paper?

The workflow writes "not reported" instead of guessing. Fabricating proof techniques, parameter signs, or equilibrium results is explicitly prohibited, and missing arXiv or Semantic Scholar metadata is not treated as a failure.

Can theory-ingest handle agent-based simulation models?

Pure computational or agent-based models are treated as experiment-like and routed to experiments pages, not propositions. The propositions directory is reserved for proven formal results only.