novelty-check

Verify research idea novelty against recent literature via multi-source search and cross-model review.

Updated Aug 18, 2026
One-click install
npx skills add https://github.com/Nitrogen216/awesome-dsh-mods --skill novelty-check-nitrogen216
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: novelty-check
Source: https://github.com/Nitrogen216/awesome-dsh-mods/tree/main/modes/dsh-autoresearch/vendor/aris-upstream/skills/novelty-check
Command: npx skills add https://github.com/Nitrogen216/awesome-dsh-mods --skill novelty-check-nitrogen216

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Before investing months into a research project, you need to know whether your proposed method has already been published. This Skill systematically checks a research idea against recent literature so you avoid pursuing work that reviewers would reject as non-novel. ## Core Features & Use Cases - Claim Decomposition: Extracts 3-5 core technical claims from a method description and evaluates each one individually for novelty. - Multi-Source Literature Search: Searches arXiv, Google Scholar, Semantic Scholar, and recent top venues (ICLR, NeurIPS, ICML) with multiple query formulations per claim. - Cross-Model Verification: Sends a novelty dossier to an external reviewer model via Codex MCP for an independent second opinion, with anti-hallucination verification of every cited paper. - Use Case: You have an idea for a new attention mechanism. Run a novelty check to receive a structured report scoring each claim, listing the closest prior work in a comparison table, and giving a PROCEED / PROCEED WITH CAUTION / ABANDON recommendation with positioning advice. ## Quick Start Ask the agent to run a novelty check on your proposed method, for example by saying "check novelty of my idea: using retrieval-augmented pruning to compress mixture-of-experts models".

Frequently Asked Questions about novelty-check

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

FAQPage Schema
How do I check if my research idea is novel?▼

Provide a description of your proposed method and the skill decomposes it into 3-5 core technical claims, then searches arXiv, Google Scholar, and Semantic Scholar for each claim. It outputs a structured report scoring each claim's novelty and listing the closest prior work.

What literature sources does a novelty check cover?▼

The search covers arXiv preprints from 2024-2026, Google Scholar, Semantic Scholar, and recent top venues including ICLR 2025/2026, NeurIPS 2025, and ICML 2025/2026. Each claim is searched with at least three different query formulations.

How does the skill avoid hallucinated paper citations?▼

Every paper in the closest-prior-work table must pass pre-search verification through a verify_papers.py helper that resolves entries against arXiv, CrossRef, and Semantic Scholar. Unverifiable entries are tagged [UNVERIFIED] rather than silently dropped or fabricated.

Can I use the novelty check without the Codex MCP reviewer?▼

The cross-model verification phase calls an OpenAI model such as gpt-5.6-sol via Codex MCP, so that integration must be available for the full workflow. The literature search and report phases rely on WebSearch, WebFetch, Grep, Read, and Glob.

What are the limitations of automated novelty checking?▼

The check depends on what is indexed and retrievable online, so very recent unpublished work or paywalled content may be missed. It also treats simple "apply X to Y" combinations as non-novel unless the application reveals surprising insights.