| name | aris-novelty-check |
| description | Verify research idea novelty against recent literature. Use when user says "novelty check", "check novelty", "has anyone done this", or wants to verify a research idea is novel before implementing. Ported from wanshuiyin/Auto-claude-code-research-in-sleep (ARIS) — methodology only, Claude-Code-specific MCP deps stripped for VS Code Copilot. |
Novelty Check Skill
Check whether a proposed method/idea has already been done in the literature.
The user will provide the method or idea description as part of their request.
Instructions
Given a method description, systematically verify its novelty:
Phase A: Extract Key Claims
- Read the user's method description
- Identify 3-5 core technical claims that would need to be novel:
- What is the method?
- What problem does it solve?
- What is the mechanism?
- What makes it different from obvious baselines?
Phase B: Multi-Source Literature Search
For EACH core claim, search using all available web sources:
-
Web Search:
- Search arXiv, Google Scholar, Semantic Scholar
- Use specific technical terms from the claim
- Try at least 3 different query formulations per claim
- Include year filters for recent work (last 2 years)
-
Known paper databases: Check against:
- Recent conference proceedings (ICLR, NeurIPS, ICML, ACL, CVPR, etc.)
- Recent arXiv preprints
-
Read abstracts: For each potentially overlapping paper, fetch its abstract and related work section
Phase C: Self-Review Verification
After gathering the literature, perform a critical self-review of novelty:
-
Write a dossier (e.g., NOVELTY_DOSSIER.md) containing:
- The proposed method description
- All papers found in Phase B
- The question: "Is this method novel? What is the closest prior work? What is the delta?"
-
Review the dossier critically, acting as a skeptical reviewer:
- For each claim, identify the closest prior work
- Determine the actual delta (what is genuinely new vs. incremental)
- Consider whether a reviewer would cite the prior work as making this redundant
Phase D: Novelty Report
Output a structured report:
## Novelty Check Report
### Proposed Method
[1-2 sentence description]
### Core Claims
1. [Claim 1] — Novelty: HIGH/MEDIUM/LOW — Closest: [paper]
2. [Claim 2] — Novelty: HIGH/MEDIUM/LOW — Closest: [paper]
...
### Closest Prior Work
| Paper | Year | Venue | Overlap | Key Difference |
|-------|------|-------|---------|----------------|
### Overall Novelty Assessment
- Score: X/10
- Recommendation: PROCEED / PROCEED WITH CAUTION / ABANDON
- Key differentiator: [what makes this unique, if any]
- Risk: [what a reviewer would cite as prior work]
### Suggested Positioning
[How to frame the contribution to maximize novelty perception]
Important Rules
- Be BRUTALLY honest — false novelty claims waste months of research time
- "Applying X to Y" is NOT novel unless the application reveals surprising insights
- Check both the method AND the experimental setting for novelty
- If the method is not novel but the FINDING would be, say so explicitly
- Always check the most recent 6 months of arXiv — the field moves fast
- Anti-hallucination for Closest Prior Work. Never fabricate arXiv IDs, DOIs, or titles from memory. Every paper in the prior-work table must be verified via web search. If a paper cannot be verified, tag it
[UNVERIFIED] and surface the uncertainty rather than dropping it.