What is the Technical Debt Conventions skill?

It is a foundational framework that provides standardized rules, schemas, and scoring rubrics for Claude agents to identify and report technical debt consistently across a project.

What are the primary debt categories defined in this skill?

The skill defines five core categories: AI Patterns, Complexity, Duplication, Architecture, and Security-related technical debt.

What safety features are included in the conventions?

It mandates content fencing to prevent prompt injection from scanned files and enforces strict path validation rules to prevent directory traversal attacks.

Does this skill perform code scanning directly?

No, it is a reference skill that defines the standards, safety rules, and output formats that other specialized scanner agents must follow to ensure interoperability.

How does it handle debt specific to AI-generated code?

It includes specific detection heuristics for 'AI patterns' such as excessive comment-to-code ratios, repeated boilerplate blocks, and generic variable naming conventions.

Technical Debt Conventions

Name: Technical Debt Conventions
Author: KingInYellows

byKingInYellows

0•

開発者ツール

Establishes a standardized framework for technical debt assessment, scoring rubrics, and automated scanner reporting patterns.

This skill serves as the foundational source of truth for managing technical debt within a development ecosystem by defining rigorous scoring rubrics, effort estimation models, and JSON output schemas. It ensures consistency across various automated scanner agents by providing clear definitions for categories such as architectural debt, AI-generated code patterns, and security risks. By implementing this skill, Claude can perform structured audits, validate scanner outputs, and generate actionable remediation plans while maintaining strict safety boundaries and path validation rules to protect the integrity of the codebase.

主な機能

01Categorized debt definitions for AI patterns, complexity, and architecture

02Built-in safety rules and content fencing to prevent prompt injection

03Multi-tier severity rubric for consistent P1-P4 issue prioritization

04Comprehensive effort estimation framework from quick fixes to large refactors

05Standardized JSON output schema (v1.0) for automated code scanners

060 GitHub stars

ユースケース

01Implementing safety guardrails for agents interacting with potentially malicious or untrusted source code

02Defining the reporting structure and validation for custom AI-driven code analysis agents

03Standardizing technical debt scoring and categorization across large, multi-language codebases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kinginyellows/yellow-plugins debt-conventions

For use in Claude.ai and ChatGPT

Download Skill