What is a PRP in the context of this skill?

A PRP (Product Requirement Prompt) is a detailed specification used within the blueprint-plugin workflow to guide Claude Code through implementation tasks with specific context and requirements.

What happens if a score is below 7?

If a score is below 7, the skill provides a 'Needs Refinement' assessment with a checklist of specific gaps to address and recommended next steps to improve the score.

How does the scoring determine delegation readiness?

The skill uses a 1-10 scale; a score of 9+ indicates the task is ready for autonomous subagent delegation, while 7-8 is suitable for direct execution with minor discovery needed.

Can I use this for existing work-orders?

Yes, you can activate this skill to review any existing work-order or PRP to identify missing context, vague implementation paths, or missing validation steps.

Does it work with different programming languages?

Yes, Confidence Scoring is language-agnostic and focuses on the structural quality and completeness of the requirements and validation logic.

Confidence Scoring

Name: Confidence Scoring
Author: laurigates

bylaurigates

•

생산성 및 워크플로

Evaluates the readiness and quality of development task definitions through systematic confidence scoring to optimize autonomous execution.

The Confidence Scoring skill provides a rigorous framework for assessing Product Requirement Prompts (PRPs) and work-orders before implementation begins. By analyzing five critical dimensions—context completeness, implementation clarity, documented pitfalls, validation coverage, and test coverage—it helps developers determine whether a task is ready for autonomous execution or requires further refinement. This skill is essential for complex workflows involving subagent delegation, ensuring that the AI has all the necessary information and executable validation gates to succeed without manual intervention.

주요 기능

013 GitHub stars

02Systematic scoring of context completeness and implementation clarity

03Standardized response templates for high and low confidence scenarios

04Validation gate verification ensuring all commands are executable and specific

05Automated readiness assessment for subagent delegation (9+ score threshold)

06Comprehensive checklists for documenting library-specific pitfalls and mitigations

사용 사례

01Determining if a generated PRP is complete enough for a subagent to handle autonomously

02Identifying specific gaps in documentation or context before starting a complex refactor

03Standardizing quality gates and validation commands across a development team's work-orders

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add laurigates/claude-plugins confidence-scoring

For use in Claude.ai and ChatGPT

Download Skill