skills.rest
microsoft
microsoftOFFICIAL
Explore all skills by this author →
Website𝕏@OpenAtMicrosoftGitHub

eval-result-interpreter

Analyzes Copilot Studio evaluation CSV results to generate a structured triage report.

123|20|Updated Mar 30, 2026
Search Tags:#playbook#csv#triage#root-cause#eval#copilot-studio
One-click install
npx skills add https://github.com/microsoft/eval-guide --skill eval-result-interpreter
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: eval-result-interpreter
Source: https://github.com/microsoft/eval-guide/tree/main/skills/eval-result-interpreter
Command: npx skills add https://github.com/microsoft/eval-guide --skill eval-result-interpreter

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The skill analyzes Copilot Studio evaluation CSV results to produce a structured triage report, including a SHIP/ITERATE/BLOCK verdict, root-cause analysis, and prioritized remediation guidance.

Core Features & Use Cases

  • Parse Copilot Studio CSV results and compute per-method, per-category, and overall pass/fail metrics.
  • Generate a SHIP / ITERATE / BLOCK verdict using the Microsoft Triage & Improvement Playbook.
  • Provide failure triage with root-cause classification, diagnostic questions, remediation mapping, and pattern analysis.
  • Output a clear next-run plan and, where possible, scenario mappings to guide improvements.
  • Accept CSV input or plain-text summaries for flexible triage.

Quick Start

Analyze Copilot Studio evaluation CSV results to generate a structured triage report with a verdict and remediation plan.

Frequently Asked Questions about eval-result-interpreter

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

FAQPage Schema
How do I analyze Copilot Studio evaluation CSV results for root-cause triage?▼

To analyze Copilot Studio evaluation CSV results, you can use a triage skill that parses the file, aggregates pass and fail metrics per method and category, and outputs a structured verdict with root-cause classification.

What is a SHIP, ITERATE, or BLOCK verdict in Copilot Studio eval triage?▼

A SHIP, ITERATE, or BLOCK verdict is a triage decision applied to Copilot Studio evaluation results to determine whether the copilot is ready for release, needs further refinement, or has critical failures requiring immediate fixes.

What columns are expected in a Copilot Studio CSV file for failure triage?▼

A Copilot Studio CSV file for failure triage should include columns such as question, expectedResponse, actualResponse, testMethodType, result, and explanation to map failures to root causes and remediation actions.

Can I use plain-text summaries instead of CSV files to generate a remediation plan?▼

Yes, you can use plain-text summaries instead of CSV files to generate a remediation plan, as the triage process accepts flexible input formats to provide root-cause analysis and prioritized actions.

Does the triage process provide a next-run plan for failed Copilot Studio test scenarios?▼

Yes, the triage process provides a next-run plan for failed Copilot Studio test scenarios by mapping failures to specific scenario improvements and offering prioritized remediation guidance for subsequent evaluations.

Related Skills

Explore More Data & Analytics Skills→
microsoftmicrosoft
123

eval-triage-and-improvement

Triage Copilot Studio eval results into structured reports with owners and actions.

Official
Advanced
telihateliha

Specification Contradiction Checker

Detect specification contradictions and inconsistencies across markdown documents.

Official
Intermediate
davidrrowleydavidrrowley

m365-agents-ts

Build multichannel Microsoft 365 agent applications with TypeScript and Express.

Community
Intermediate
polsegupolsegu

test-flakiness

Detect nondeterministic failures by parsing CI logs and aggregating per-test pass/fail histories.

Community
Intermediate