consult-sofia

Delegate read-only diagnostic tasks to the Coval Sofia expert system.

2|Updated Feb 17, 2026
One-click install
npx skills add https://github.com/coval-ai/coval-external-skills --skill consult-sofia
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: consult-sofia
Source: https://github.com/coval-ai/coval-external-skills/tree/main/skills/sofia/consult-sofia
Command: npx skills add https://github.com/coval-ai/coval-external-skills --skill consult-sofia

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill bridges the gap between raw evaluation data and actionable insights by delegating complex diagnostic tasks to Sofia, an expert system specialized in Coval FDE analysis.

Core Features & Use Cases

  • Diagnostic Analysis: Provides expert-level interpretation of failed runs, simulation regressions, and monitoring conversations.
  • Strategic Recommendations: Offers grounded advice on test coverage, persona design, and metric configuration based on your organization's specific data.
  • Use Case: When a voice agent evaluation shows a sudden drop in performance, use this skill to analyze the specific run IDs and receive a prioritized plan for agent improvement and test case adjustment.

Quick Start

Use the consult-sofia skill to analyze the recent regression in run_123 and propose a targeted fix for the agent.

Frequently Asked Questions about consult-sofia

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

FAQPage Schema
How do I analyze a failed voice agent evaluation run?

To analyze a sudden drop in agent performance, provide the specific run IDs to the Coval Sofia expert system. It interprets the simulation regressions and monitoring conversations to deliver grounded, FDE-style recommendations for targeted fixes.

What is FDE analysis in the context of simulation diagnostics?

FDE analysis is the expert interpretation of organization-specific evaluation metrics, run failures, and agent performance regressions. It bridges raw evaluation data and actionable insights by providing grounded advice on test coverage, persona design, and metric configuration.

Do I need an active Coval API connection to run diagnostic analysis?

Yes, you need an active Coval API connection and a valid organization context. The diagnostic analysis requires this environment to access your evaluation data and provide grounded, FDE-style recommendations.

How do I get strategic recommendations for test coverage and persona design?

You can get strategic recommendations for test coverage and persona design by delegating read-only analytical tasks to the Coval Sofia expert system. It evaluates your organization-specific metrics to offer grounded configuration advice.

Can I use this approach to fix agent performance regressions?

Yes, you can fix agent performance regressions by delegating the specific run IDs to the expert system. It provides a prioritized plan for agent improvement and targeted test case adjustments based on the diagnostic findings.