A professional-grade observability suite for monitoring AI agent tool invocations, LLM latency, and operational costs in real-time.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install claw-reliability
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 claw-reliability 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 Claw Reliability?
Claw Reliability is a robust observability skill designed for the Openclaw ecosystem, providing deep insights into agent behavior and system health. It serves as a dedicated monitoring layer that tracks every interaction between your AI agent and its environment, ensuring that developers can maintain high uptime and performance standards.
By integrating this tool from the Openclaw Skills collection, you gain access to a localized SQLite-backed engine that parses gateway events and session transcripts. It offers a transparent view of token consumption, tool success rates, and potential logic loops, making it indispensable for production-grade AI deployments.
Claw Reliability Use Cases
- Tracking real-time token usage and projected API costs for LLM calls.
- Identifying bottlenecks in agent performance by analyzing tool invocation latency.
- Detecting infinite loops or repetitive tool failures before they consume excessive resources.
- Auditing agent history for security by flagging unusual or first-time tool usage.
- Visualizing session health through a dedicated React-based dashboard.
How Claw Reliability Works
- The monitoring daemon attaches to the Openclaw gateway to listen for session events and tool execution logs.
- Raw events are processed and normalized into a structured schema within a local SQLite database.
- The skill analyzes incoming data against predefined thresholds to identify cost spikes or tool failures.
- If anomalies are detected, the system triggers pluggable alerts to destinations like Slack or Discord.
- Developers can query this data via CLI commands or view it through the integrated FastAPI and React web dashboard.
Claw Reliability Setup
Ensure you have python3 installed and your workspace configured. Install the skill and start the monitoring daemon:
# Start the monitoring service
cd {baseDir} && python3 scripts/monitor.py start --config {baseDir}/config.yaml
# Configure alert destinations (Optional)
cd {baseDir} && python3 scripts/monitor.py configure-alerts --destination discord --webhook-url <URL>
# Launch the visual dashboard
cd {baseDir} && python3 dashboard/backend/main.py
Claw Reliability Data Schema & Taxonomy
All observability data is stored locally in {baseDir}/data/metrics.db. The schema includes:
| Table | Description |
|---|---|
| Tool Invocations | Tracks tool name, success/fail status, duration, and arguments. |
| LLM Calls | Logs model metadata, token counts (in/out), latency, and cost. |
| Session Data | Records lifecycle events, message counts, and timestamps. |
| Alert Logs | Stores historical records of anomalies and severity levels. |
Claw Reliability Advanced Features
- Real-time Anomaly Detection: Automatic identification of cost spikes and tool loops.
- Pluggable Alerting: Integration with Discord, Slack, and local log files for instant notifications.
- Local-First Privacy: All metrics remain in your local environment; no data is sent externally by default.
- Customizable Thresholds: Fine-tune failure rates and usage limits via the configuration YAML.
- Interactive Visualizations: A full-stack React dashboard for high-level session overviews.
SKILL.md
Loading
Related Openclaw Skills
Face Transform
A comprehensive AI-powered facial transformation suite for analysis, editing, and stylistic modification of portrait images.
Suno Music Generation
A professional AI music production skill that leverages the Suno API to generate full songs, lyrics, and stems from simple text prompts.
doc-extract-filter
A versatile document processing tool for high-performance text extraction, multi-level filtering, and batch file management with OCR support.
Automated Marketing Manager
An AI-driven marketing engine designed to automate content creation, post scheduling, and engagement tracking across TikTok, Instagram, and Telegram.
Audit System Skill
A specialized inspection framework for identifying logic errors, security risks, and UX friction in automated systems.
Bilibili Transcriber
A professional Bilibili video processing agent that transcribes audio into text and generates high-quality structured summaries.
METADATA
Featured*
