> ## Documentation Index > Fetch the complete documentation index at: https://critiqor-71f5274a.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Critiqor Dashboard: Agent Reliability Report Guide > The Critiqor dashboard is a local visual report generated after each observation session. Learn how to open it, navigate it, and act on what it shows. The Critiqor dashboard is a local, visual reliability report generated at the end of each observation session. It surfaces a trust-level verdict, a primary diagnosis, supporting evidence, and concrete recommendations — all scoped to a single run of your agent. This page covers how to open the dashboard, how it's laid out, and where to go first. Critiqor dashboard preview showing a healthy run with a 100 trust score, active runs, recent runs, and primary diagnoses panels.

Critiqor dashboard preview showing a healthy run with a 100 trust score, active runs, recent runs, and primary diagnoses panels.

## Opening the dashboard You can open the dashboard in three ways depending on what you need: **After finalizing a session** — the dashboard opens automatically when you run: ```bash theme={null} critiqor finalize ``` **Most recent run** — to reopen the last generated report at any time: ```bash theme={null} critiqor dashboard ``` **A specific run** — to open the report for a particular session by its run ID: ```bash theme={null} critiqor dashboard run_001 ``` ## Dashboard layout The dashboard is organized into three broad areas: * **Top — Run Summary:** High-level signals for the session, including the Executive Summary verdict and Trust Score. Start here to get a fast read on overall health. * **Middle — Diagnostic Detail:** The Primary Diagnosis, Runtime Evidence, Runtime Timeline, and Recommendations sections. Go here when you need to understand what happened and what to do about it. * **Bottom — History & Settings:** The Historical Runs section, which lists previous sessions so you can track reliability over time, and Settings for managing how the current run is stored. ## Quick navigation The top-level verdict for your run. Start here to decide whether to dig deeper. A reliability level (High, Moderate, or Low) backed by observed confidence signals. The main finding for the run. Your entry point for debugging when something went wrong. The detailed record of tool calls, retries, failures, and behavioral patterns observed during the session. A chronological view of agent activity. Use it to spot retry storms, gaps, and chains of dependent failures. Actionable next steps tied to specific findings. Your improvement checklist after reviewing a run. A list of previous sessions with trust levels and diagnosis summaries for tracking trends over time. Configuration options for the current run, including how it is stored and whether it appears in history. Start with the **Executive Summary** and **Trust Score** to get a quick read on the run. Then move to **Primary Diagnosis** and **Runtime Evidence** when you need the full detail behind a finding. The dashboard runs entirely on your local machine. Nothing is sent to external servers unless you have explicitly configured a backend integration.