A professional-grade research agent that enforces data integrity through mandatory source carding and evidence-based reporting.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install deep-research-v21
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 deep-research-v21 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 Deep Research Pro v2.1?
Deep Research Pro is a high-fidelity research framework designed for Openclaw Skills users who require absolute precision and verifiable data. Unlike standard AI search tools, this skill enforces a strict no-source-no-report policy, ensuring every claim in your final output is backed by a specific, quality-scored evidence card. It is built to prevent AI hallucinations by mandating intermediate steps, including structured research planning and cross-analysis of conflicting data points.
The system utilizes advanced protocols to fetch academic papers via the OpenAlex API and industry insights through web browsing. Version 2.1 introduces enforced depth, requiring a minimum of 15 validated source cards before a final report can even be generated. This makes it an essential tool for analysts, academics, and developers who need to synthesize complex information without sacrificing technical truth.
Deep Research Pro v2.1 Use Cases
- Academic literature reviews requiring strict DOI/URL citations and P-value extraction.
- Industry market analysis where cross-referencing conflicting data points is critical.
- Technical due diligence for software projects and emerging technologies.
- Automated evidence gathering for whitepapers and professional reports.
How Deep Research Pro v2.1 Works
- Research Planning: The agent generates a comprehensive
research/plan.mdfile with at least 5 targeted query strings and waits for user approval. - Source Carding: The agent searches for sources and immediately generates individual
sources/card-xxx.mdfiles for every valid find, extracting quotes and core data. - Quality Scoring: Each source is passed through a scoring script to ensure it meets the minimum authority and relevance threshold.
- Cross-Analysis: The agent synthesizes the gathered cards into
analysis/synthesis.md, specifically highlighting at least 3 areas of contradictory data. - Report Generation: A final report is compiled only if the 15-card minimum is met, with mandatory
[[card-xxx]]citations for every data point.
Deep Research Pro v2.1 Setup
To initialize this Openclaw Skills research module, ensure you have the necessary API access and file permissions configured:
# Verify your environment has the quality score script
ls scripts/quality-score.py
# Run the research command to start a deep dive
/research "Your Topic Here" --depth deep
# Check source card count during the process
ls sources/card-*.md | wc -l
Deep Research Pro v2.1 Data Schema & Taxonomy
The skill maintains a strict directory structure to ensure data provenance:
| Directory | File Type | Purpose |
|---|---|---|
research/ |
plan.md |
Initial search queries and strategy. |
sources/ |
card-001.md... |
Individual evidence blocks with quotes and metrics. |
analysis/ |
synthesis.md |
Comparative analysis and conflict resolution. |
reports/ |
final-report.md |
The final cited output with verified links. |
Deep Research Pro v2.1 Advanced Features
- Enforced Depth Mode: Automatically rejects report generation if fewer than 15 high-quality sources are found.
- Conflict Attribution: Mandatory identification of data discrepancies with attribution to specific methodology or regional differences.
- Automated Quality Gates: Integrated Python scripts to score source reliability on a 0-10 scale.
- Persistent Context: Retains source cards even if a process is interrupted, allowing for resume-based workflows in Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
PDF OCR Extractor
A local, privacy-focused tool that converts image-based PDFs into searchable text using Tesseract OCR and Python.
Page Saver
A local, zero-dependency reference tool providing instant access to Page Saver documentation, patterns, and best practices directly in your terminal.
Kiwi Receipts
A dedicated New Zealand tax assistant that transforms receipt photos into IRD-ready GST reports and comprehensive annual tax summaries.
Save to Email
A powerful utility to send formatted HTML emails and notifications through the Resend API via a simple shell interface.
Disney Plus Information Organizer
A specialized tool for extracting and organizing public metadata, membership details, and creator guidelines from the Disney Plus platform.
Chinese Sentence Correction
A precision-engineered AI skill for identifying and correcting grammatical and logical errors in Chinese text while strictly preserving original meaning.
Featured*
