_evaluation

Coordinate multi-perspective evaluation workflows with parallel evaluators and threshold filtering.

Updated Mar 28, 2026
One-click install
npx skills add https://github.com/HahyeonJeon/gobbi --skill evaluation-hahyeonjeon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: _evaluation
Source: https://github.com/HahyeonJeon/gobbi/tree/main/plugins/gobbi/skills/_evaluation
Command: npx skills add https://github.com/HahyeonJeon/gobbi --skill evaluation-hahyeonjeon

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Evaluation coordination in ClaudeX workflows requires multi-perspective judgment and governance to avoid bias and ensure user control.

Core Features & Use Cases

  • Orchestrates perspective selection and parallel evaluators
  • Aggregates independent verdicts and routes findings to the user
  • Supports confidence/severity scoring, threshold filtering, and stage-aware criteria

Quick Start

Load Evaluation Perspective Quality at the start of your workflow to coordinate multi-perspective evaluation.

Frequently Asked Questions about _evaluation

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

FAQPage Schema
How do I coordinate multi-perspective evaluation workflows to avoid bias?

Multi-perspective evaluation coordination selects diverse perspectives and spawns separate evaluators with task-specific criteria, aggregating independent verdicts to avoid bias and routing findings for user decision making.

How does an evaluation orchestrator apply stage-appropriate criteria?

An evaluation orchestrator applies stage-aware criteria by assigning task-specific rules to spawned evaluators at each workflow stage, then filtering aggregated results through confidence and severity scoring thresholds.

Can I enforce threshold filtering on evaluation verdicts with confidence and severity scores?

Yes, evaluation workflows support scoring with confidence and severity, enforce threshold filtering on aggregated verdicts, and route only findings meeting defined criteria to the user.

What is the best way to aggregate independent evaluation verdicts from parallel evaluators?

Aggregating independent verdicts requires an orchestrator to coordinate parallel evaluators, collect their separate judgments, apply stage-appropriate criteria, and consolidate findings for routing to the user.

When do I need governance in an evaluation workflow?

Evaluation governance is needed when workflows require multi-perspective judgment and user control to avoid bias, ensuring independent verdicts are routed to the user for final decision making.

Does the evaluation orchestrator provide integration guidance for reporting?

Yes, the evaluation orchestrator provides integration guidance for reporting after aggregating independent verdicts and enforcing threshold filtering on confidence and severity scores.