/xbaseline

description: Establish and track quality, performance, and security baselines with regression detection
tags: [baseline, metrics, performance, quality, regression, thresholds]

Baseline Management

Establish intelligent baselines for quality, performance, and security metrics. Detect regressions against baselines. Aligns with the Baseline Management pattern from ai-development-patterns.

Usage Examples

**Capture current baseline:**

/xbaseline capture

**Compare current state against baseline:**

/xbaseline check

**Show baseline drift over time:**

/xbaseline drift

**Set custom thresholds:**

/xbaseline threshold --coverage 85 --complexity 8

**Help and options:**

/xbaseline help
/xbaseline --help

Implementation

If $ARGUMENTS contains "help" or "--help": Display this usage information and exit.

Step 1: Detect Project Context

Examine project for measurable metrics: !ls -la | grep -E "(pyproject.toml|package.json|go.mod|Cargo.toml)" !find . -name ".coverage" -o -name "coverage.xml" -o -name "lcov.info" | head -5 !find . -name "*.py" -o -name "*.js" -o -name "*.ts" -o -name "*.go" | grep -v node_modules | wc -l

Step 2: Execute Based on Mode

**Mode 1: Capture Baseline (argument: "capture")** If $ARGUMENTS contains "capture":

Collect current metrics and save as baseline:

1. **Code Metrics**: !find . -name "*.py" -not -path "*/node_modules/*" -not -path "*/__pycache__/*" | xargs wc -l 2>/dev/null | tail -1 !find . -name "*.py" -not -path "*/node_modules/*" | wc -l 2>/dev/null

Measure:

• Total lines of code
• Number of source files
• Average file length
• Functions exceeding 25 lines
• Cyclomatic complexity distribution

2. **Test Metrics**: !python -m pytest --co -q 2>/dev/null | tail -1 || echo "Test count not available"

• Number of tests
• Test execution time
• Coverage percentage

3. **Quality Metrics**: !ruff check . --statistics 2>/dev/null | tail -5 || echo "No linter baseline"

• Linting violations count
• Type errors count
• TODO/FIXME count

!grep -r "TODO\|FIXME\|HACK" . --include="*.py" --include="*.js" --include="*.ts" | wc -l 2>/dev/null

4. **Security Metrics**:

• Known dependency vulnerabilities
• Hardcoded secret patterns found
• Dangerous code patterns count

5. **Dependency Metrics**:

• Total dependency count
• Outdated dependencies
• Dependencies with known CVEs

Save baseline to `.baseline.json` in project root with timestamp.

**Mode 2: Check Against Baseline (argument: "check" or no arguments)** If $ARGUMENTS is empty or contains "check":

Compare current state against saved baseline:

1. Read `.baseline.json` (if not found, suggest running `capture` first) 2. Collect current metrics using same methods 3. Compare each metric and flag regressions:

• Coverage dropped below baseline
• Lint violations increased
• Complexity increased
• New security issues introduced
• Test count decreased (tests deleted?)

Report as:

• IMPROVED: Metric better than baseline
• STABLE: Metric within acceptable range
• REGRESSED: Metric worse than baseline (flag with details)

**Mode 3: Drift Analysis (argument: "drift")** If $ARGUMENTS contains "drift":

Show how metrics have changed since baseline was captured:

• Read `.baseline.json` for historical reference
• Calculate delta for each metric
• Identify trends (improving, stable, degrading)
• Highlight areas needing attention

**Mode 4: Custom Thresholds (argument: "threshold")** If $ARGUMENTS contains "threshold":

Override default thresholds. Parse remaining arguments for:

• `--coverage N` — minimum coverage percentage
• `--complexity N` — maximum cyclomatic complexity
• `--violations N` — maximum lint violations
• `--file-length N` — maximum file length

Save custom thresholds to `.baseline.json`.

Baseline Report Format

## Baseline Report — [date]

| Metric | Baseline | Current | Status |
|--------|----------|---------|--------|
| Test coverage | 82% | 79% | REGRESSED |
| Lint violations | 12 | 8 | IMPROVED |
| Avg complexity | 4.2 | 4.5 | STABLE |
| File count | 45 | 48 | STABLE |
| TODO count | 8 | 15 | REGRESSED |

### Regressions Requiring Action
- Coverage dropped 3% — review recently added code without tests
- TODO count increased — schedule tech debt cleanup

Keep output focused on regressions and actionable improvements.