What is the Confidence Check skill?

It is a specialized tool for Claude Code that runs a 5-point assessment to ensure an implementation plan is valid, unique, and architecturally sound before any code is written.

What does the skill check for specifically?

It assesses five criteria: absence of duplicates (25%), architecture compliance (25%), official documentation verification (20%), working OSS references (15%), and root cause identification (15%).

How does this skill save tokens?

By spending a small amount of tokens on a pre-check (100-200), you avoid the massive token cost (5,000-50,000) of generating and then fixing an incorrect or redundant implementation.

Why is a 90% confidence score required?

The 90% threshold ensures that the AI has verified the root cause, checked for duplicates, and aligned with project docs, minimizing the risk of 'hallucinated' solutions or technical debt.

Confidence Check

Name: Confidence Check
Author: Activer007

byActiver007

0•

部署与 DevOps

Validates implementation readiness through a multi-point assessment to prevent wasted effort and architectural drift.

The Confidence Check skill acts as a pre-implementation gatekeeper for AI-driven development, ensuring that tasks are executed with high precision and minimal waste. It mandates a 90% confidence threshold before any coding begins, achieved by systematically verifying the absence of duplicate code, confirming alignment with the project's specific architecture, and cross-referencing official documentation. By identifying the root cause and finding proven OSS implementations first, this skill significantly reduces token consumption and prevents the AI from heading in the wrong technical direction.

主要功能

01Quantitative scoring system (0.0 - 1.0) with a 90% pass threshold

02Architectural alignment checks against CLAUDE.md and PLANNING.md

03Automated duplicate detection using Grep and Glob patterns

04Root cause identification to ensure fixes target the source of issues

050 GitHub stars

06Real-time official documentation and OSS reference verification

使用场景

01Scanning large repositories to prevent the creation of redundant utility functions or modules.

02Ensuring a proposed feature implementation aligns with the existing tech stack and design patterns.

03Verifying a bug fix strategy before writing code to ensure the root cause is properly identified.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add activer007/ordinary-claude-skills confidence-check

For use in Claude.ai and ChatGPT

Download Skill