A high-precision performance monitoring tool that provides millisecond-level measurement of the AI agent response lifecycle.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install response-speed-test
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 response-speed-test 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 Response Speed Test?
The Response Speed Test is a specialized diagnostic tool designed for developers working with Openclaw Skills to monitor and optimize system latency. It provides a granular breakdown of the time spent in every phase of an agent's operation, from the initial gateway reception to the final message delivery, ensuring that developers can pinpoint exact bottlenecks in their workflows.
By providing detailed metrics such as Time To First Token (TTFT) and Tokens Per Second (TPS), this skill allows for objective comparisons between different LLM providers and configuration settings. It is essential for anyone building production-grade agents where response speed and user experience are critical priorities.
Response Speed Test Use Cases
- Benchmarking different LLM models and providers within the Openclaw Skills framework.
- Identifying latency bottlenecks in memory loading or session dispatching phases.
- Continuous monitoring of system health via automated heartbeat-triggered tests.
- Generating professional performance reports for stakeholders or technical audits.
How Response Speed Test Works
- The user triggers a measurement via a single run command or a multi-iteration benchmark.
- The skill attaches timestamps to key lifecycle events, starting from the moment the user message is sent.
- It monitors internal Openclaw Skills events, including gateway receiving, session dispatching, and memory retrieval.
- During the LLM interaction, it specifically captures the start time and the arrival of the very first token to calculate TTFT.
- Upon completion, the skill calculates the total response time and token generation speed (TPS).
- All data is synthesized into a detailed report or terminal output for analysis.
Response Speed Test Setup
Install the skill using the following command:
clawhub install response-speed-test
Run a basic test to verify performance:
response-speed-test run
Response Speed Test Data Schema & Taxonomy
The skill organizes performance data across eight distinct stages of the message lifecycle:
| Stage | Event Code | Description |
|---|---|---|
| T0 | MESSAGE_SENT | Initial user message dispatch |
| T1 | GATEWAY_RECEIVE | Gateway successfully receives the payload |
| T2 | SESSION_DISPATCH | The session is allocated and dispatched |
| T3 | MEMORY_LOAD | Contextual memory is loaded into the session |
| T4 | LLM_REQUEST_START | The request is sent to the LLM provider |
| T5 | LLM_FIRST_TOKEN | Time To First Token (TTFT) is recorded |
| T6 | LLM_RESPONSE_COMPLETE | LLM finishes generating the response |
| T7 | MESSAGE_SENT | The final processed message is delivered |
Key Metrics tracked: TTFT (Time To First Token), TPS (Tokens Per Second), and Total Latency.
Response Speed Test Advanced Features
- Iterative Benchmarking: Run multiple tests sequentially using
--iterationsto calculate average performance and stability. - Automated Scheduling: Configure periodic runs within
HEARTBEAT.mdto ensure your Openclaw Skills maintain peak performance over time. - Multi-Format Reporting: Export detailed performance analysis in Markdown or other formats for integration with documentation.
- Deep Hook Integration: Accesses low-level system events to ensure millisecond-level precision in measurement.
SKILL.md
Loading
Related Openclaw Skills
VT Sentinel Security Plugin
Integrate real-time VirusTotal malware scanning and threat protection into your AI agent workflow.
Quality-Driven Development
A sophisticated development methodology skill that enforces TDD, DDD, and the TRUST 5 quality framework to produce production-ready code.
OpenClaw Server Security & Hardening
A professional-grade security framework for hardening Linux servers and deploying protected OpenClaw agent instances.
PLVR Event Discovery
Automate the discovery and comparison of live events on plvr.io using browser automation.
Security Audit Skill
A comprehensive security auditing toolkit for scanning codebases, infrastructure, and agentic AI systems for vulnerabilities and misconfigurations.
Bazi Chart Generator
A specialized tool for generating traditional Chinese Bazi charts and 10-step Great Luck cycles based on birth data.
Featured*
