How does this skill help in saving LLM tokens?

By spending approximately 100-200 tokens on a pre-implementation check, it prevents the potential waste of 5,000 to 50,000 tokens on incorrect, redundant, or hallucinated code.

What is the minimum confidence score required to proceed?

Implementation only begins if the skill calculates a confidence score of 0.90 (90%) or higher based on its five-point check system.

What factors are used to calculate the confidence score?

The score is derived from five checks: No Duplicate Implementations (25%), Architecture Compliance (25%), Official Documentation Verified (20%), Working OSS References (15%), and Root Cause Identified (15%).

Can this skill be used for debugging?

Yes, it is highly effective for debugging as it includes a specific check for root cause identification to ensure the AI understands the actual problem before suggesting a fix.

Confidence Check

Name: Confidence Check
Author: Microck

byMicrock

•

보안 및 테스팅

Assesses implementation readiness via a multi-factor confidence score to prevent redundant or incorrect code execution.

The Confidence Check skill acts as a pre-implementation safeguard that ensures Claude has verified all necessary prerequisites before writing code. By calculating a weighted confidence score based on duplicate detection, architecture alignment, official documentation review, open-source references, and root cause identification, it prevents 'wrong-direction' development. This skill enforces a strict 90% confidence threshold, significantly reducing token wastage and implementation errors by ensuring the AI fully understands the technical context and existing codebase patterns before proceeding.

주요 기능

0181 GitHub stars

02Real-time official documentation and OSS reference validation

03Architecture and tech stack compliance verification

04Automated duplicate implementation detection via Grep and Glob

05Weighted multi-factor confidence scoring (0.0 to 1.0)

06Root cause analysis for bug fixes and error-driven tasks

사용 사례

01Ensuring new modules align with established architectural patterns defined in CLAUDE.md

02Verifying if a proposed feature already exists in a large codebase to avoid technical debt

03Validating API usage against live documentation before implementing third-party integrations

What are Skills?·How to Install

Install with 🐟 Skill.Fish

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

For use in Claude.ai and ChatGPT

Download Skill