What is weighted voting in this context?

Weighted voting assigns higher influence to agents with relevant domain expertise. For example, a security agent's vote counts more heavily on encryption tasks than a general-purpose agent's vote.

How does the confidence multiplier work?

Each agent's weighted vote is multiplied by their confidence level (High = 1.0, Medium = 0.7, Low = 0.4), ensuring that uncertain agents have less impact on the final decision.

Can I use this for general code generation?

No, this skill is best used for discrete decision-making. For parallel code generation and implementation, the n-version-workflow is recommended instead.

What happens if agents don't reach a consensus?

If the configured threshold (like a 66% supermajority) isn't met, the skill reports the failure and suggests escalating to a 'debate-workflow' for deeper analysis.

When does this skill auto-activate?

It triggers automatically when Claude detects changes involving security implementations, authentication logic, encryption, or sensitive data handling.

Consensus Voting

Name: Consensus Voting
Author: rysweet

byrysweet

•

Security & Testing

Executes weighted multi-agent voting to validate critical technical decisions and security implementations through specialized expertise.

The Consensus Voting skill facilitates structured decision-making by leveraging multiple specialized agents with domain-weighted expertise. Designed for high-stakes scenarios like security architectures, authentication changes, and complex data handling, it ensures that the most qualified AI agents guide the final outcome. By applying weighted scores—such as giving a security agent triple influence on encryption choices—the skill provides a transparent audit trail, confidence metrics, and supermajority validation to reduce the risk of architectural flaws in production code.

Key Features

01Intelligent domain detection to select the most relevant agents for every decision

02Domain-weighted voting that prioritizes specialized expertise over simple majorities

03Detailed consensus reporting featuring weighted scores, confidence levels, and dissenting views

0417 GitHub stars

05Configurable voting thresholds including simple majority, supermajority, and unanimous

06Automatic triggers for high-risk tasks like security, encryption, and authentication

Use Cases

01Selecting between cryptographic hashing algorithms like Argon2id vs BCrypt for user authentication

02Validating architectural shifts in sensitive data handling or PII management to ensure compliance

03Reaching multi-agent consensus on complex API rate limiting or access control strategies

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rysweet/amplihack consensus-voting

For use in Claude.ai and ChatGPT

Download Skill