Glossary
Codebase glossary
Short, sourced definitions of the terms used to judge a codebase: complexity, churn, hotspots, bus factor, blast radius, dead code, decision records and the agent context behind them. Each entry says how the number is measured, how to measure it yourself with free tools, and how repowise measures it.
Last reviewed: November 2026
Code health
Code health
Code health measures how costly the next change to a file will be, as a score from 1 to 10. What goes into it, how to read it and how to check it against bugs.
Code smell
A code smell is a surface sign of a deeper design problem. Fowler and Beck's classic smells, with examples, free detection tools and how repowise flags them.
Cognitive complexity
Cognitive complexity scores how hard code is to read: +1 per break in linear flow plus a nesting penalty. Rules, a worked example and the cyclomatic difference.
Cyclomatic complexity
Cyclomatic complexity counts the independent paths through code: M = E - N + 2P, or decision points plus 1. Formula, a worked example and common thresholds.
Technical debt
What is technical debt? The meaning, Ward Cunningham's origin, Fowler's four-quadrant model, how to measure it and which debt to pay down first.
Git history
Bus factor
Bus factor is the number of people who would have to leave before a project or file stalls. How it is computed from git history, with a worked example.
Code churn
Code churn in software development is how much code is added, changed or deleted in a file over a period. How to measure it with git, with a worked example.
Code hotspot
A code hotspot is a file that changes often and is hard to change safely. How churn and complexity find hotspots with git, plus a free command to start.
Code ownership
Code ownership is who is responsible for, and knows, a piece of code. How owners are measured from git blame and commits, and how it differs from CODEOWNERS.
Temporal coupling
Temporal coupling (co-change, logical coupling) is when files keep changing in the same commits. How support and confidence measure it, with a worked example.
Change and test risk
Blast radius
Blast radius meaning in software: everything a code change or failure can affect. How to measure it from the call graph, with a worked example.
Change risk
Change risk analysis scores a commit or pull request for defect risk from its size, spread and history. How it works, with free ways to measure it.
CRAP score
The CRAP score combines cyclomatic complexity and test coverage to flag risky, untested code. See the formula, a worked table and the usual threshold of 30.
Test coverage gap
A test coverage gap is code that matters or keeps changing but no test runs. Patch vs total coverage, how to find gaps, and the report formats repowise reads.
Structure and dead code
Call graph
A call graph maps which functions call which. Static vs dynamic call graphs, why dynamic dispatch makes them hard, free tools, and how repowise scores edges.
Dead code
Dead code is code that never runs or that nothing uses. Unreachable vs unused code, how to find it with free tools per language, and how to delete it safely.
Dependency graph
A dependency graph shows what depends on what. Package graph vs code import graph, how cycles and fan-in are measured, and free tools such as madge and pydeps.
Zombie dependency
A zombie dependency is a declared package that nothing imports. How it differs from an unused import, and free tools to find one in JS, Python, Rust and Go.
Architecture and agents
Architecture decision record
An architecture decision record (ADR) is a short file that records one design decision, its context and consequences. Template, tools, and how to mine them.
C4 model
The C4 model describes software architecture at four zoom levels: context, containers, components and code. What each level shows and how to generate them.
Codebase intelligence
Codebase intelligence is a kept-current model of a repository: code graph, git history, docs, decisions and health, queried by people and agents.
MCP server
An MCP server exposes tools to AI agents over the Model Context Protocol. What a codebase MCP server does, how it connects, and what to expect from one.
See every one of these on your repo.
Index a repository free and read its hotspots, owners, dead code and change risk from your own history. Rolling it out to a team? We will walk you through it.