zicato-tournament-forensics

Explain zicato tournament promotion, rejection, or deferral decisions from settled artifacts.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill explains why a specific zicato champion-versus-challenger tournament round was promoted, rejected, or deferred without treating the decision as a black box.

Core Features & Use Cases

  • Verdict Transparency: Inspect the authoritative gate decision, rejection reason, scalar delta, pass-rate delta, and margin calculation.
  • Matchup Forensics: Analyze the per-entry A/B grid, scalar breakdown, pass flips, mutation heatmap, score trajectory, and tournament cost.
  • Artifact-Based Investigation: Read experiment records, patches, loss profiles, harness provenance, and round logs to diagnose invalid or misleading verdicts.
  • Use Case: When a challenger is rejected despite improving aggregate drift loss, use this Skill to determine whether a pass-rate regression, namespace monotonicity violation, holdout failure, or untested mutation caused the decision.

Quick Start

Use the zicato tournament forensics skill to explain why the specified parent and child generation matchup was promoted, rejected, or deferred using its settled workspace artifacts.

Frequently Asked Questions about zicato-tournament-forensics

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

FAQPage Schema
How do I explain why a tournament round promoted or failed without rerunning evaluations?

Diagnose tournament promotion decisions by inspecting settled artifacts, gate outcomes, and runtime telemetry. This approach reveals rejection reasons like pass-rate regressions or holdout failures without spending live evaluation budget.

Why does a challenger get rejected despite improving aggregate drift loss?

A challenger can be rejected despite lower drift loss due to pass-rate regressions, namespace monotonicity violations, holdout failures, or untested mutations. Inspect the scalar breakdown and mutation heatmap to identify the exact cause.

How do I analyze per-entry regressions in a champion-versus-challenger matchup?

Analyze per-entry regressions by examining the A/B grid, pass flips, and score trajectory from settled round artifacts. This isolates specific entries causing the verdict without rerunning the tournament.

Can I diagnose invalid tournament verdicts using only experiment records and patches?

Yes, you can diagnose invalid or misleading verdicts by reading experiment records, patches, loss profiles, and harness provenance. These artifacts provide sufficient data to validate the gate decision without executing new evaluations.

What is mutation validation in generation comparison tournaments?

Mutation validation checks whether code changes between generations are properly tested and applied. It uses heatmaps and per-entry matchups to ensure untested or unexpected mutations do not cause misleading promotion gate decisions.

Does tournament forensics require live budget-spending evaluations to inspect gate outcomes?

No, tournament forensics inspects gate outcomes, scalar deltas, pass-rate monotonicity, and holdout confirmation directly from settled round artifacts. It bypasses live budget-spending evaluations entirely by using existing runtime telemetry.