Council Reference Data FAQs

Question 1

How are false positives filtered out?

Accepted Answer

The skill includes a specific taxonomy that auto-rejects findings if they are pre-existing issues, pedantic nitpicks, or items already caught by standard linters and compilers.

Question 2

What is the mandatory response format for findings?

Accepted Answer

All council agents must return a structured JSON object including the consultant name, confidence score, severity level, specific file location, and actionable recommendations.

Question 3

Can I invoke this skill directly in my terminal?

Accepted Answer

No, this skill is configured as non-invocable and is intended to serve as background reference data for other subagents within the Claude Code environment.

Question 4

How does the expertise weight matrix work?

Accepted Answer

It assigns specific coefficients to different AI models (such as Gemini or Codex) based on their proven performance in specific domains like Security, Performance, or Debugging.

Question 5

What is the purpose of the Council Reference skill?

Accepted Answer

It provides the standardized data and logic rules for a multi-agent review system, ensuring different AI models provide consistent, structured, and high-quality feedback.

Council Reference Data

주요 기능

사용 사례

Council Reference Data

주요 기능

사용 사례