zicato-watch-dashboard

Monitor zicato's live Console dashboard and interpret multi-agent harness evolution runs.

4|2|Updated May 14, 2026
One-click install
npx skills add https://github.com/pedapudi/zicato --skill zicato-watch-dashboard
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: zicato-watch-dashboard
Source: https://github.com/pedapudi/zicato/tree/main/skills/zicato-watch-dashboard
Command: npx skills add https://github.com/pedapudi/zicato --skill zicato-watch-dashboard

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps operators understand whether zicato's self-improvement loop is running, healthy, and actually reducing loss without manually piecing together telemetry, tournament state, and evaluation results.

Core Features & Use Cases

  • Live Dashboard Monitoring: Open and navigate the local Console across environments, epochs, rounds, candidates, boards, evaluations, mutations, traces, and publications.
  • Tournament Interpretation: Read racing funnels, Swiss ladders, elimination flows, and gauntlet delta lanes to identify survivors, eliminations, promotions, and champion defenses.
  • Run Health and Liveness: Distinguish fresh in-flight activity from stale heartbeats, inspect active board units, and detect toothless evaluations that cannot reliably distinguish candidates.
  • Evidence-Based Retrospectives: Compare loss-floor trends, champion timelines, board heatmaps, transcripts, mutation diffs, and epoch reports to determine whether the harness improved.
  • Use Case: During or after an evolve run, use this Skill to report the current dashboard URL, inspect the active tournament, explain what is running, and summarize whether the latest challenger displaced the champion.

Quick Start

Use the zicato watch dashboard skill to inspect the current local Console and summarize the active epoch, tournament progress, evaluation health, and improvement outcome.

Frequently Asked Questions about zicato-watch-dashboard

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

FAQPage Schema
How do I monitor multi-agent system runs and check evaluation health in real time?

To monitor multi-agent system runs, you need a dashboard monitoring tool that reads live Console telemetry, heartbeat freshness, and file-backed JSON endpoints to report evaluation health without starting a new evolve run.

How do I interpret tournament analysis results like Swiss ladders and gauntlet delta lanes?

Interpreting tournament analysis requires reading racing funnels, Swiss ladders, and gauntlet delta lanes to identify candidate survivors, eliminations, promotions, and champion defenses across in-flight and post-mortem epochs.

What is the best way to check if a self-improvement loop is actually reducing loss?

Checking if a self-improvement loop reduces loss involves comparing loss-floor trends, champion timelines, and board heatmaps to conduct evidence-based retrospectives on whether the harness improved.

Does dashboard monitoring support post-mortem epoch analysis without triggering a live evolve run?

Dashboard monitoring supports post-mortem epoch analysis by inspecting file-backed JSON endpoints, mutation diffs, transcripts, and publication reports to summarize run health without starting a live evolve run.

Why does my multi-agent dashboard show stale heartbeats during an active tournament round?

Stale heartbeats in a multi-agent dashboard indicate a lack of fresh in-flight activity, requiring inspection of active board units and evaluation health to detect toothless evaluations that cannot distinguish candidates.

Can I inspect candidate mutations and champion timelines for post-mortem analysis?

You can inspect candidate mutations and champion timelines by navigating the local Console to compare mutation diffs, epoch reports, and transcripts, determining whether the latest challenger displaced the champion.