Quality Analysis

CodexA provides a full code quality pipeline with metrics tracking, hotspot detection, impact analysis, and CI-ready quality gates.

Quality Analysis

Run a comprehensive scan with:

bash

codex quality src/
codex quality --json            # JSON output
codex quality --safety-only     # Security-only (fast)

What Gets Analyzed

Category	Tool	Description
Complexity	Radon	Cyclomatic complexity per function
Security	Bandit	Common vulnerability patterns
Dead Code	Built-in	Unused functions, unreachable code
Duplication	Built-in	Copy-paste detection
Maintainability	Built-in	Composite MI score (0–100)

Maintainability Index

A per-file and project-wide score (0–100) based on:

Lines of code — penalizes overly large files
Cyclomatic complexity — penalizes deeply nested logic
Comment ratio — rewards well-documented code

MI Range	Rating
65–100	Good (easy to maintain)
40–64	Moderate
0–39	Poor (difficult to maintain)

Hotspot Detection

Find the riskiest code in your project:

bash

codex hotspots
codex hotspots --top-n 10 --json
codex hotspots --no-git       # Skip git churn data

Risk Factors

Factor	Weight	Source
Complexity	0.30	Cyclomatic complexity
Duplication	0.20	Duplicate line density
Fan-in	0.15	Number of callers
Fan-out	0.15	Number of callees
Churn	0.20	Git change frequency

When git data is unavailable, churn weight is redistributed across the other factors.

Impact Analysis

Predict the blast radius of changes:

bash

codex impact parse_file          # Analyze a function
codex impact src/parser.py       # Analyze a file
codex impact MyClass --max-depth 3 --json

Impact analysis uses BFS over the call graph and dependency map to find:

Direct dependents — Functions that call the target
Transitive dependents — Functions affected indirectly
Affected modules — Files that import the target
Dependency chains — Full paths from target to affected code

Symbol Trace

Trace execution flow through the codebase:

bash

codex trace parse_file
codex trace MyClass.process --json

Shows upstream callers and downstream callees to map how execution flows.

Metrics & Trends

Track quality over time:

bash

codex metrics                  # Current metrics
codex metrics --snapshot       # Save a snapshot
codex metrics --history 10     # Last 10 snapshots
codex metrics --trend          # Direction analysis

Trends report: improving, stable, or degrading for each metric.

Quality Gates

Enforce quality policies in CI:

bash

codex gate --strict            # Exit code 1 on failure
codex gate --min-maintainability 60 --max-complexity 15

Policy	Default	Description
`min_maintainability`	40.0	Minimum MI score
`max_complexity`	25	Maximum cyclomatic complexity
`max_issues`	20	Maximum total issues
`max_dead_code`	15	Maximum dead code symbols
`max_duplicates`	10	Maximum duplicate pairs
`require_safety_pass`	true	Safety check must pass

CI Integration

Generate CI workflow templates:

bash

codex ci-gen analysis           # Full analysis workflow
codex ci-gen safety             # Lightweight safety-only
codex ci-gen precommit          # Pre-commit hook config

Plugin Hooks

Quality analysis fires these hooks for customization:

Hook	When
`PRE_HOTSPOT_ANALYSIS`	Before hotspot scoring
`POST_HOTSPOT_ANALYSIS`	After hotspot report
`PRE_IMPACT_ANALYSIS`	Before impact BFS
`POST_IMPACT_ANALYSIS`	After impact report
`PRE_TRACE`	Before symbol trace
`POST_TRACE`	After trace result

Quality Analysis ​