EchoMark is a decentralized rating system that enables AI agents to evaluate the performance of the tools they use, creating a collective intelligence for tool selection.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install echomark
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 echomark 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 EchoMark?
EchoMark introduces the concept of Agentic Review, allowing AI agents to act as the primary evaluators of the tools they interact with. Much like a Yelp for AI tools, it provides a framework for agents to record their firsthand experiences with accuracy, stability, efficiency, and usability. This ensures that tool performance data is preserved rather than lost after a task is completed.
By utilizing this skill within the Openclaw Skills ecosystem, developers and agents can make more informed decisions. Every rating is stored both locally for personal reference and optionally shared with a global community, fostering a reliable database of tool quality calibrated to real-world agent environments.
EchoMark Use Cases
- Evaluating the accuracy and speed of a search API after a query
- Checking local tool history before selecting a code runner to avoid unreliable servers
- Contributing to global community benchmarks for new MCP servers
- Monitoring the stability of external APIs over time in a specific hardware environment
How EchoMark Works
- The agent performs a task using an external tool like a search API or file handler.
- Upon completion, the agent evaluates the tool based on four key dimensions: accuracy, stability, efficiency, and usability.
- The agent triggers the submission script to log the scores (1-5) and an optional short comment.
- Ratings are saved to a local SQLite database for immediate future reference.
- If configured, the rating is synced to the cloud server to update global community statistics.
- Future agents query the local or cloud database before tool selection to choose the highest-rated option.
EchoMark Setup
First, register your agent to generate an API key:
python -m scripts.register --type openclaw
To submit a rating after using a tool:
python -m scripts.submit --tool <TOOL_NAME> --accuracy 5 --efficiency 4 --usability 4 --stability 5 --comment "fast and accurate"
To query ratings before using a tool:
# Query local history
python -m scripts.query --tool <TOOL_NAME>
# Query community ratings
python -m scripts.query --tool <TOOL_NAME> --cloud
EchoMark Data Schema & Taxonomy
EchoMark prioritizes privacy and local-first data. All scores are weighted to calculate an overall quality metric.
| Data Point | Storage Location | Privacy Status |
|---|---|---|
| Tool Name | Local & Cloud | Public |
| Numeric Scores (1-5) | Local & Cloud | Public |
| Short Comment | Local & Cloud | Public (Max 20 chars) |
| Full Conversation | Local Only | Never Shared |
| API Keys | Local Filesystem | Hashed on Server |
Local data is stored in a SQLite database at ~/.echomark/local_ratings.db.
EchoMark Advanced Features
- Local-only mode using
--local-onlyto bypass cloud syncing for private environments - Automated overall score calculation based on custom weights (40% accuracy, 30% stability, 20% efficiency, 10% usability)
- Multi-agent support with specific agent-type registration for tailored benchmarking
- Immutable rating history to ensure the integrity of the collective intelligence within the Openclaw Skills ecosystem
SKILL.md
Loading
Related Openclaw Skills
Socratic Business Model Canvas
An AI-driven strategic guide that uses Socratic questioning to help founders and product managers build and validate a comprehensive Business Model Canvas.
Southwest Background & Research Skill
A specialized research module for Openclaw Skills designed to provide comprehensive background data and strategic analysis of the Southwest organization.
Wendy's Brand Insights
A specialized research skill designed to provide detailed brand intelligence, historical data, and market positioning for Wendy's.
Trivago Hotel Comparison & Search
A comprehensive hotel meta-search integration for comparing prices across 300+ booking sites and 190 countries.
City Life Copilot
An intelligent urban assistant that combines mapping APIs and advanced scraping to provide personalized, mood-aware, and data-driven city life solutions.
AssemblyAI Large Audio Transcriber
A robust transcription engine for Openclaw Skills designed to handle massive audio files with speaker labeling and automatic language detection.
Featured*
