agentforce-observe

What problem does it solve?

This Skill helps you diagnose why an Agentforce agent behaved incorrectly by analyzing production session traces, preview traces, routing decisions, action usage, and instruction adherence. It turns scattered telemetry into clear root causes for regressions, performance issues, and knowledge gaps.

Core Features & Use Cases

• Production observability: Review STDM session data, conversation turns, tool calls, and quality signals from live orgs.
• Local reproduction: Recreate suspected failures in preview sessions and inspect trace files for routing or grounding problems.
• Root-cause analysis: Detect subagent misroutes, missing actions, wrong bindings, dead hubs, abandoned sessions, and LOW instruction adherence.
• Fix guidance: Map symptoms back to the agent configuration and recommend targeted edits to the authoring bundle.
• Use case: An AI product analyst can use this Skill to understand why an agent failed in production, confirm the issue in preview, and prepare the next fix.

Quick Start

Ask the skill to analyze an Agentforce agent’s recent session traces and summarize the main failures, root causes, and recommended fixes.