An academic research verification tool that extracts claims from meeting records and validates novelty against a 1.2-billion-document database.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install novelty-validator
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install novelty-validator using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Novelty Validator?
The Novelty Validator is an academic research verification tool designed to check if ideas, methodologies, or research directions discussed in meetings have already been explored in existing literature. By converting transcripts and text descriptions into structured claims, this tool matches ideas against a 1.2-billion-document database to locate potential overlaps.
Operating within the ecosystem of Openclaw Skills, the validator integrates with literature databases to identify similar academic works and patents. It is built to support researchers during brainstorming phases, group meetings, and advisor sessions, offering a quick method to check research uniqueness without manual database cross-referencing.
Novelty Validator Use Cases
- Academic Group Seminars: Verify discussed ideas immediately during or after a meeting to confirm they have not been previously published.
- Advisor Consultation Preparation: Check the novelty of your proposed thesis topics before discussing them with your advisor.
- Patent Verification Prep: Analyze engineering or application-oriented ideas against patent databases to confirm baseline uniqueness.
- Cross-Lab Collaboration: Validate newly formulated hypotheses during joint academic sessions using structured meeting transcripts.
How Novelty Validator Works
- Idea Collection: The tool pulls meeting transcripts using the tmeet command-line tool or accepts direct text inputs detailing a research idea.
- Structured Claim Extraction: The raw input is parsed into a structured format containing the problem, method, and claimed novelty.
- Multi-Engine Academic Search: The extracted claim is run across parallel databases including Chinese journals, global databases, and patent registries via global-biblio-base.
- Vector Similarity Re-ranking: The system calculates semantic distances using vector embeddings to identify conceptual similarities, even when phrasing differs.
- Multi-Agent Adversarial Debate: Roles such as Searcher, Skeptic, Defender, and Arbiter debate the claim against prior work to identify overlapping areas.
- Assessment & Reporting: A report is compiled classifying the idea from Level 1 (Fully Covered) to Level 5 (Novel), accompanied by verifiable URLs and DOIs.
Novelty Validator Setup
To deploy the Novelty Validator within the Openclaw Skills environment, configure your Tencent Meeting CLI access.
Prerequisite: Tencent Meeting CLI Authentication
Ensure your meeting integration tool is authenticated by checking its status:
tmeet auth status
If the tool is unauthenticated, initiate the login sequence to retrieve the browser authorization link:
tmeet auth login 2>&1 &
Once logged in, you can pull transcripts from specific sessions using your meeting identifiers:
tmeet record list --meeting-id <meeting_id> --compact
Novelty Validator Data Schema & Taxonomy
The Novelty Validator processes information through structured definitions to maintain trace evidence.
Claim Definition Schema
| Attribute | Format | Description |
|---|---|---|
Problem |
String | The core research question or problem statement to resolve |
Method |
String | The technological or theoretical framework proposed |
Claimed Novelty |
String | The specific aspects of the methodology believed to be unique |
Five-Level Collision Matrix
- Level 1 (Fully Covered): Overlap across all core attributes; idea has been published.
- Level 2 (Highly Similar): Overlap in three core attributes; high similarity.
- Level 3 (Partially Overlapping): Overlap in two core attributes; requires differentiation.
- Level 4 (Adjacent): Overlap in only one attribute; indicates incremental advancement.
- Level 5 (Novel): No substantial overlap detected in existing databases.
Novelty Validator Advanced Features
- Three-Engine Parallel Execution: Queries global literature, regional databases, and international patent databases to secure wide search coverage.
- Multi-Agent Debate Framework: Employs independent reasoning agents to stress-test claims, minimizing subjective biases during evaluation.
- Embedded Semantic Mapping: Automatically falls back to localized word-overlap metrics if external vector computation endpoints are unreachable, maintaining stable execution inside Openclaw Skills.
- Verifiable Evidence Trails: Provides direct links, DOIs, and original abstract fragments for every flagged literature conflict, ensuring all decisions remain verifiable.
SKILL.md
Loading
Related Openclaw Skills
Agent Handshake
Openclaw Skills turns AI servers into discoverable, persistent agents with one-command registration, acceptance, and handshake workflows.
Agent Deployment Kit
A structured methodology to replicate an AI agent configuration across multiple clients using a single context source and a strict, validation-gated generator.
Canvora Visual Generation Skill
Canvora is a CLI-driven utility that converts raw text, web URLs, and documents into high-quality, on-brand marketing visuals and structured slide decks.
Tristan RFQ Overseer
Tristan RFQ Overseer turns inbound RFQs into priced, compared, and approval-gated quotation drafts with vault-backed state.
Academic Defense Rehearsal
An Openclaw Skills workflow that rehearses academic defenses end-to-end with mock committees, literature weak-spot scans, and delivery debriefs.
Stock Planner
Stock Planner turns a holdings CSV into a next-day trading plan with P/L analysis, stop-loss and take-profit checks, and exposure control.
Featured*
