observability

Generate semantic metadata JSON with toxicity assessment from SKILL.md files.

1|Updated Dec 18, 2025
One-click install
npx skills add https://github.com/matt-fell-ai/PAI --skill observability-matt-fell-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: observability
Source: https://github.com/matt-fell-ai/PAI/tree/main/skills/Observability
Command: npx skills add https://github.com/matt-fell-ai/PAI --skill observability-matt-fell-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires bun, express, vite, vue, and includes scripts (resource) and apps (resource) components.

What problem does it solve?

Stop wondering what your AI agents are doing. This Skill provides live WebSocket streaming of agent sessions, tool calls, and execution timelines.

Core Features & Use Cases

  • Agent Activity Monitoring: Track session starts/ends, tool calls across all agents, and hook event execution with comprehensive visibility into multi-agent workflows.

Core Features & Use Cases

  • Real-Time Tracking: Monitor agent sessions, hook events, and WebSocket live updates.
  • Use Case: Imagine your multi-agent workflow seems stuck. Use this Skill to see real-time tool calls and identify which agent is blocked or looping.
  • Strategic Debugging: Quickly identify bottlenecks, performance issues, and coordination problems in complex AI operations.

Quick Start

Start the observability dashboard to monitor all agent activity in real-time via WebSocket streams.

Core Features & Use Cases

  • Live Dashboard: Access the web interface at http://localhost:5172 for immediate insight into all active PAI operations.

Frequently Asked Questions about observability

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I monitor agent activity in real-time across a multi-agent workflow?

Real-time agent monitoring tracks session starts, tool calls, and hook events through WebSocket streams. Access the live dashboard at http://localhost:5172 to observe all active PAI operations and agent coordination as it happens.

How can I debug a stuck or looping multi-agent workflow?

Strategic debugging identifies bottlenecks and coordination problems by displaying live tool calls and execution timelines. Watch which agent is blocked or looping in real-time to pinpoint performance issues quickly.

What's the best way to track tool execution across multiple agents?

Comprehensive visibility into multi-agent workflows comes from monitoring tool calls across all agents with hook event execution data. The dashboard streams execution timelines so you see the full call sequence and agent interactions.

Can I use this monitoring dashboard with my existing Express and Vue stack?

Yes, the observability Skill integrates with Express backends and Vue frontends via WebSocket streaming. It works within your existing tech stack to provide live dashboard updates without replacing core components.

What visibility does WebSocket streaming provide for agent sessions?

WebSocket streaming delivers live updates on session starts, session ends, tool calls, and hook event execution. This continuous data stream enables immediate detection of workflow delays and agent state changes.