code-complexity

name	code-complexity
description	Analyse code for complexity issues and suggest reductions. Use when the user asks to check complexity, simplify code, or reduce technical debt.

Evaluate the code for complexity using these signals:

Cyclomatic complexity — count decision points (if, else, switch cases, loops, catch, ternaries, logical operators). Flag any function with more than ~10 branches; suggest splitting or extracting early returns.

Cognitive complexity — focus on nesting depth and logical operator density. Deeply nested blocks are harder to read than equivalent flat code; flatten with guard clauses or extracted functions.

Coupling — identify modules that reach into each other's internals, share mutable state, or have implicit execution ordering. High coupling makes isolated changes risky.

Dead code — flag unused variables, unreachable branches, obsolete classes or interfaces. Dead code adds noise without value.

Size — functions or files that do too much. A function that needs a paragraph to describe what it does is a candidate to split.

For each issue found:

• Name the specific metric or signal
• Show the problematic code span
• Give a concrete refactor suggestion (early return, extracted function, interface boundary, deletion)

Severity guide:

• Must fix: cyclomatic complexity > 15, coupling that makes testing impossible, widespread dead code
• Should fix: nesting > 3 levels, functions > ~40 lines, high Halstead vocabulary
• Consider: moderate coupling, mid-size functions that could be cleaner