An automated 5-dimension benchmarking system for AI agents to self-evaluate reasoning, retrieval, creation, execution, and orchestration capabilities.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install botlearn-assessment
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 botlearn-assessment 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 BotLearn Capability Assessment?
The BotLearn Capability Assessment is a sophisticated evaluation framework designed as one of the premier Openclaw Skills for measuring AI performance. It functions simultaneously as an exam administrator and an examinee, conducting a rigorous 5-dimension test covering reasoning, retrieval, content creation, execution, and tool orchestration.
By leveraging this skill, developers can obtain an objective, data-driven benchmark of an agent's true operational capacity. The system uses a randomized question pool across varying difficulty levels to ensure a comprehensive and unbiased analysis of how well an agent handles complex workflows and tool dependencies.
BotLearn Capability Assessment Use Cases
- Benchmarking new AI model integrations within an agent workflow to ensure quality standards.
- Periodic self-evaluation of agent capabilities after system updates or configuration changes.
- Generating detailed performance reports including radar charts for stakeholder reviews and technical audits.
- Auditing tool-use proficiency and orchestration logic in complex, multi-tool environments.
How BotLearn Capability Assessment Works
- Identify the exam intent through natural language triggers such as benchmark me or run assessment.
- Select random questions across five core dimensions (Reasoning, Retrieval, Creation, Execution, Orchestration) based on difficulty levels.
- Execute tasks autonomously, outputting the question text first followed by the agent response to maintain transparency for the invigilator.
- Perform a self-evaluation using a structured scoring rubric and a correction factor to ensure objective results.
- Generate comprehensive reports in Markdown and HTML formats, including a visual SVG radar chart of the final scores.
BotLearn Capability Assessment Setup
To integrate this assessment into your environment, ensure you have the core Openclaw Skills infrastructure ready. The reporting features require Node.js for radar chart generation.
# Navigate to your skill directory
cd path/to/botlearn-assessment
# Ensure node.js is available for report scripts
node --version
# Trigger the assessment via your agent interface
# Example: "run full assessment"
BotLearn Capability Assessment Data Schema & Taxonomy
The skill organizes evaluation data within a structured results directory to maintain a clear history of performance. All Openclaw Skills data is stored using the following schema:
| File | Description |
|---|---|
| exam-{sessionId}-data.json | Raw structured assessment data and metadata. |
| exam-{sessionId}-{mode}.md | Human-readable Markdown summary report. |
| exam-{sessionId}-report.html | Interactive HTML report with embedded radar charts. |
| exam-{sessionId}-radar.svg | Standalone SVG visualization of the 5-dimension scores. |
| INDEX.md | A historical index of all previous assessment runs for trend tracking. |
BotLearn Capability Assessment Advanced Features
- Automatic language adaptation based on the user's initial trigger message.
- Tool-dependency auto-detection that gracefully fails and skips tests if required CLI tools are unavailable.
- Weighted scoring algorithms that prioritize high-impact dimensions like Reasoning (25%) and Execution (20%).
- Historical trend analysis and comparison capabilities within the View History mode of the Openclaw Skills suite.
SKILL.md
Loading
Related Openclaw Skills
Feishu Card Sender Skill
A professional-grade solution for sending interactive Feishu card messages directly via OpenAPI within the Openclaw Skills ecosystem.
Ryanair Fare Finder
A specialized tool for programmatically generating and interpreting Ryanair fare-finder URLs to discover the cheapest budget flights.
BaseCred Onchain Reputation
A unified onchain reputation profiler that aggregates scores from Ethos, Talent Protocol, and Farcaster for any 0x address.
Basecred ERC-8004 Registration
An interactive tool for registering AI agents on the ERC-8004 on-chain registry using a guided, automated chat workflow.
OpenClaw Examiner
A professional assessment framework that benchmarks AI agent proficiency across eight core dimensions using standardized testing and radar chart reporting.
OpenClaw Graduate Coach
A specialized coaching skill that transforms early AI agent adoption into a long-term growth strategy through retrospective analysis and archetype discovery.
METADATA
Featured*
