agentforce-observe
Find failures in your live Agentforce conversations.
What it does
Diagnoses production Agentforce failures from session evidence, reproduces confirmed issues, guides fixes, and manages monitoring alerts.
When to use it
Use for production agent failures, session traces, performance investigations, issue reproduction, or Agent Health Monitoring alert management—not routine development or test-spec work.
How to use it
Give it an authenticated org alias and a Salesforce project; it reports production issues or manages alerts, asking whether to reproduce or improve findings.
What you provide
- the org alias
- An existing project
Uses
Good to know
- cannot be undone: Agent Health Monitoring
What you need · 5
Requires the Salesforce CLI, with declared version requirement >=2.136.8.
Requires Python >=3.10.0 for the supplied JSON-parsing commands; Windows instructions substitute python for python3.
Access · 4
This skill
Agent Health Monitoring
Creates, updates, and deletes alerts (irreversible)
Salesforce
Reads production conversations and metrics