Can I run the review on specific parts of my project?

Yes, you can run targeted reviews by providing a specific directory path, or use category filters like '--category=deps' to focus only on specific types of debt.

How is the Debt Score calculated?

The score starts at 100 and deducts points based on the severity of issues found. Critical issues subtract 10 points, while high, medium, and low issues subtract 5, 2, and 1 points respectively, with caps per category.

Where are the audit results saved?

Full audits are saved as JSON snapshots in the '.claude/reviews/' directory, allowing the skill to track trends and improvements over time.

Does this skill modify my source code?

No, the Tech Debt Review skill is strictly read-only. It performs analysis and generates recommendations without altering your files.

Tech Debt Reviewer

Name: Tech Debt Reviewer
Author: Obsidian-Owl

byObsidian-Owl

0•

Security & Testing

Analyzes codebases for technical debt, security vulnerabilities, and maintainability issues using ten specialized AI agents.

The Tech Debt Reviewer is a comprehensive diagnostic tool for Claude Code that orchestrates ten specialized agents to perform deep-dive audits of software projects. It evaluates critical metrics including cyclomatic complexity, dead code, dependency security, documentation gaps, and architectural smells to provide a quantified 'Debt Score' (0-100). By synthesizing findings into a prioritized, actionable report with trend tracking, it helps development teams maintain high code quality, reduce maintenance burdens, and visualize project health over time without modifying a single line of code.

Key Features

01Prioritized action plans focusing on high-risk, low-effort fixes

02Trend tracking with historical comparisons stored in local review snapshots

030 GitHub stars

04Flexible execution modes including PR-only, full-audit, and targeted directory scans

05Multi-agent orchestration for complexity, dead code, and dependency analysis

06Automated Debt Scoring based on issue severity and remediation impact

Use Cases

01Security and dependency reviews to catch vulnerable packages and outdated libraries

02Periodic codebase health checks to identify aging TODOs and architectural decay

03Pre-PR quality gates to ensure new changes don't introduce maintainability regressions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add obsidian-owl/floe tech-debt-review

For use in Claude.ai and ChatGPT