lc-from-paper

Reproduce a published scientific paper as an ASTRA project from a DOI, arXiv ID, or PDF.

18|2|Updated Feb 15, 2026
One-click install
npx skills add https://github.com/LightconeResearch/lightcone-cli --skill lc-from-paper
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lc-from-paper
Source: https://github.com/LightconeResearch/lightcone-cli/tree/main/claude/lightcone/skills/lc-from-paper
Command: npx skills add https://github.com/LightconeResearch/lightcone-cli --skill lc-from-paper

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill helps you reproduce a published scientific paper as a complete ASTRA project, coordinating paper extraction, spec construction, evidence resolution, implementation, execution, and results comparison without you having to manage every phase manually.

Core Features & Use Cases

  • End-to-end reproduction workflow: runs ORIENT to extract the paper, interview you, and scan any reference code, then executes the long middle via an autonomous ralph loop.
  • ASTRA-driven provenance and validation: produces an astra.yaml spec, resolves cited evidence with verified quotes, validates the spec, and compares reproduced outputs against paper targets.
  • Handles continuation/resume: supports resuming an in-flight reproduction workdir by skipping ORIENT when artifacts already exist.
  • Best-fit for replication tasks: use when you have a DOI, arXiv ID, or PDF (or you explicitly ask to reproduce that paper).

Quick Start

Tell the agent which paper to reproduce by running lc-from-paper on the paper identifier (DOI, arXiv ID, or PDF path) from your analysis workdir.

Frequently Asked Questions about lc-from-paper

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

FAQPage Schema
How do I reproduce a scientific paper end-to-end from a DOI or arXiv ID?

Paper reproduction from a DOI or arXiv ID runs an automated workflow that extracts the publication, generates a specification, resolves evidence, implements code, and compares reproduced outputs against original targets.

Can I resume an in-flight paper reproduction workdir without starting over?

You can resume paper reproduction by skipping the initial orientation phase when artifacts already exist, allowing the workflow to continue the long middle execution loop from an existing workdir.

How does research provenance and evidence validation work during replication?

Research provenance validation generates an astra specification, resolves cited evidence with verified quotes, and runs validation checks to satisfy schema and evidence integrity throughout the reproduction process.

What inputs do I need to start reproducing a paper as a complete project?

To reproduce a paper as a complete project, you need a DOI, arXiv ID, or PDF path, and an analysis workdir to initiate the extraction, spec generation, and implementation workflow.

Does paper reproduction orchestration handle both code extraction and results comparison?

Paper reproduction orchestration handles reference code scanning during orientation and executes results comparison against paper targets after the autonomous implementation and execution phases complete.

What are the limitations of autonomous paper reproduction workflows?

Autonomous paper reproduction requires running an orientation phase in the main session and launching a separate loop for the long middle, meaning manual session management is still needed for phase transitions and validation.