discover

Generate ranked paper shortlists from anchor, topic, or wiki seeds.

77|16|Updated May 9, 2026

Search Tags:#literature review #deduplication #wiki #semantic scholar #ranking #research workflow #paper discovery

Ready to run

npx skills add https://github.com/Lambenthan/empiricalwiki --skill discover-lambenthan

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: discover
Source: https://github.com/Lambenthan/empiricalwiki/tree/main/.claude/skills/discover
Command: npx skills add https://github.com/Lambenthan/empiricalwiki --skill discover-lambenthan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It helps you quickly determine which papers to read next by generating a ranked shortlist based on anchors, a topic query, or your current wiki state.

Core Features & Use Cases

Ranked paper discovery: Produces an ordered shortlist with rationale so you can judge relevance before reading.
Three seed modes: Supports anchor-based discovery (ideal after ingest), topic-based exploration, and wiki-derived recommendations from recently edited pages.
Dedup against existing wiki: Filters out papers already present in your wiki to avoid wasting review time.
Non-ingesting by design: Outputs candidates only; you choose whether to ingest them afterward.

Quick Start

Use discover to get a reading shortlist anchored on a paper ID by asking: “Run /discover with --anchor 2106.09685 and --limit 10 so I can see the best next related papers with rationales.”

Frequently Asked Questions about discover

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

FAQPage Schema

How do I find related papers to read next without manually searching?▼

You can generate a ranked shortlist of candidate papers for next-reading decisions using an anchor paper ID, a topic query, or your recently edited wiki pages. The shortlist includes rationales to help you judge relevance before reading.

What is the best way to get paper recommendations based on my existing wiki?▼

To get paper recommendations based on your wiki, use wiki-derived discovery mode. It generates candidates from recently edited wiki pages while deduplicating against papers already present in your wiki to avoid wasting review time on duplicates.

Can I get paper discovery results without automatically ingesting them into my research workflow?▼

Yes, you can get paper discovery results without ingesting them. The discovery process is non-ingesting by design, outputting only machine-readable checkpoints and a user-facing rationale so you can choose whether to ingest candidates afterward.

How do I explore related work around a specific research topic?▼

To explore related work around a research topic, use topic-based exploration mode. It generates a ranked shortlist of candidate papers with rationales, allowing you to survey the direction and judge relevance before committing to reading or ingesting.

Does the paper discovery shortlist filter out duplicates from my existing wiki?▼

Yes, the paper discovery shortlist filters out duplicates from your existing wiki. It deduplicates candidates against papers already present in your wiki pages, ensuring your reading shortlist only contains new candidate papers to review.