fabrcore-agentmonitor

Captures FabrCore agent messages, events, and LLM calls for monitoring and debugging.

Updated Feb 16, 2026
One-click install
npx skills add https://github.com/vulcan365/FabrCore --skill fabrcore-agentmonitor
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fabrcore-agentmonitor
Source: https://github.com/vulcan365/FabrCore/tree/main/docs/skills/fabrcore-agentmonitor
Command: npx skills add https://github.com/vulcan365/FabrCore --skill fabrcore-agentmonitor

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Captures and centralizes visibility into FabrCore agent activity by recording messages, events, and internal LLM calls, enabling debugging, auditing, and UI monitoring.

Core Features & Use Cases

  • Pluggable monitoring API with three independent buffers for messages, events, and LLM calls
  • Real-time notifications via OnMessageRecorded, OnEventRecorded, and OnLlmCallRecorded
  • In-memory default with optional payload capture and redaction for privacy

Quick Start

Install FabrCore server, enable the in-memory agent message monitor, and start querying recent messages, events, and LLM calls to observe agent activity.

Frequently Asked Questions about fabrcore-agentmonitor

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

FAQPage Schema
How do I monitor agent traffic and capture LLM calls in real-time?

Real-time monitoring of agent traffic captures messages, events, and internal LLM calls using a pluggable monitor API with three independent buffers. It records per-agent activity and sends UI notifications via OnMessageRecorded, OnEventRecorded, and OnLlmCallRecorded.

What is the best way to debug and audit per-agent activity across development and production?

Debugging and auditing per-agent activity works across development and production scenarios by recording messages, events, and LLM calls. An in-memory default buffer centralizes visibility into agent traffic for querying recent activity.

Can I capture agent message payloads without exposing sensitive data?

Optional payload capture with redaction for privacy allows recording agent messages without exposing sensitive data. The in-memory monitor supports configurable payload capture to balance debugging visibility with privacy requirements.

Does the agent message monitor support real-time UI notifications for event streams?

Real-time UI notifications for event streams are supported through OnMessageRecorded, OnEventRecorded, and OnLlmCallRecorded callbacks. These notify UI components immediately as new agent traffic is captured in the buffers.

How do I track token usage from LLM calls made by agents?

Token tracking from LLM calls is handled by a dedicated buffer that records internal LLM calls independently from messages and events. This allows monitoring and auditing per-agent LLM usage and token consumption.

What are the limitations of the in-memory default for agent monitoring?

The in-memory default buffer stores recent agent traffic in memory, limiting historical retention to available memory capacity. For persistent auditing, reference captures via a references directory can supplement the in-memory monitoring.