Research Harness for Openclaw

An open-source prompt stack designed for comprehensive public-market research, covering everything from company deep-dives to red-teaming investment theses.

joansongjr
v0.7.0
Apr 14, 2026
0
753
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install research-harness

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install research-harness 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 Research Harness?

Research Harness is a professional-grade research framework built for equity analysts, fund managers, and retail investors. It provides a structured stack of 13 modular skills that streamline the research lifecycle across global markets, including US, Hong Kong, and A-shares. By integrating this toolkit into your workflow with Openclaw Skills, you can automate data gathering, perform rigorous model checks, and generate high-impact decision summaries for portfolio managers.

This skill transforms a standard AI agent into a sophisticated financial analyst capable of handling complex tasks like value chain mapping and earnings preview analysis. It bridges the gap between raw market data and actionable investment insights through a standardized evidence grading system and multi-source data adapters.

Research Harness Use Cases

  • Conducting a 9-section deep-dive into a new company under coverage to establish a baseline.
  • Mapping industry value chains to identify key players and supply-demand imbalances.
  • Red-teaming an existing investment thesis to uncover potential risks and falsification paths.
  • Preparing for earnings calls with automated metric tracking and consensus gap analysis.
  • Generating one-page PM Briefs for fast decision-making in high-pressure trading environments.

How Research Harness Works

  1. Installation via the Clawhub registry activates the core routing logic within the agent environment.
  2. Upon initialization, the agent displays a comprehensive 13-point menu for user selection, ranging from research to risk analysis.
  3. The user inputs a number or describes a task, which is processed by the Autopilot or Master Mode logic to route to the correct sub-skill.
  4. The skill selects the appropriate data adapter, prioritizing specialized sources like iFind MCP or global web search engines.
  5. The agent executes the specific research module using an evidence grading system (F1/F2/M1/C1/H1) to ensure technical accuracy.
  6. The final output is delivered as a structured markdown report, briefing, or decision journal entry.

Research Harness Setup

To get started with this toolkit in your Openclaw Skills environment, use the following command:

clawhub install research-harness

After installation, simply trigger the skill to see the interactive menu. For optimal performance, ensure your environment has access to the iFind MCP or the cn-web-search adapter for deep market data.

Research Harness Data Schema & Taxonomy

Research Harness utilizes several key files to maintain context and track research quality across sessions:

File Purpose
watchlist.md Tracks the active coverage universe of stocks and sectors.
biases.md Logs research biases to be checked during Red Team sessions.
decision-log.md A journal documenting the rationale behind investment decisions.

Additionally, the skill implements an Evidence Grading system to categorize data:

  • F1/F2: Fact-based evidence levels.
  • M1: Model-driven assumptions and sensitivity checks.
  • C1/H1: Consensus expectations and hypothesis tracking.

Research Harness Advanced Features

  • Autopilot and Master Mode for intelligent task routing and autonomous multi-step research execution.
  • Multi-market support tailored specifically for the nuances of A-shares, HK stocks, and US equities.
  • Dynamic data adapter decision tree that intelligently chooses between MCP, web search, and manual input based on availability.
  • Integrated Red Team module designed to act as a devil's advocate against established bullish cases.
  • Full compatibility with the Openclaw Skills ecosystem for seamless multi-agent collaboration and research automation.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*