Does this skill require external internet access?

To reach full confidence, the skill uses MCP tools like Tavily or WebFetch to verify official documentation and find working open-source implementation examples online.

What happens if the confidence score is below 90%?

If the score is between 70% and 90%, the tool suggests alternatives or asks clarifying questions; if it falls below 70%, it stops the task entirely to request more context.

How does it verify architecture compliance?

The skill scans project-specific files like CLAUDE.md and PLANNING.md to ensure the proposed solution aligns with your established tech stack and design patterns.

Why should I use the Confidence Check skill?

It acts as a safeguard that prevents Claude from starting tasks with insufficient context, which saves thousands of tokens and prevents the creation of duplicate or non-compliant code.

Confidence Check

Name: Confidence Check
Author: agiletec-inc

byagiletec-inc

0•

Productividad y Flujo de Trabajo

Validates implementation readiness by assessing technical feasibility, architectural alignment, and duplicate code checks before execution.

The Confidence Check skill is a pre-implementation gatekeeper for Claude Code designed to prevent 'wrong-direction' development and high token waste. By calculating a rigorous confidence score based on five critical metrics—duplicate detection, architectural compliance, documentation verification, OSS referencing, and root cause identification—it ensures Claude proceeds only when it has a 90% or higher certainty of success. This skill is essential for maintaining codebase integrity and optimizing development costs by catching potential logic errors or architectural mismatches before implementation begins.

Características Principales

010 GitHub stars

02Automated codebase search for duplicate implementations

03Categorized action prompts: Proceed, Investigate, or Stop

04Integration with MCP tools for official documentation and OSS research

05Architectural alignment verification against project standards

06Standardized 0.0 to 1.0 confidence scoring system

Casos de Uso

01Verifying complex architectural changes before starting a major refactor

02Validating technical feasibility against official API documentation and proven OSS examples

03Ensuring new features do not duplicate existing logic in large or unfamiliar codebases

Características Principales

010 GitHub stars

02Automated codebase search for duplicate implementations

03Categorized action prompts: Proceed, Investigate, or Stop

04Integration with MCP tools for official documentation and OSS research

05Architectural alignment verification against project standards

06Standardized 0.0 to 1.0 confidence scoring system

Casos de Uso

01Verifying complex architectural changes before starting a major refactor

02Validating technical feasibility against official API documentation and proven OSS examples

03Ensuring new features do not duplicate existing logic in large or unfamiliar codebases