review-medical-compliance

Checks medical software best practices using the medical-compliance-reviewer agent.

Usage

/review-medical-compliance [area]

Where area is:

• •logging - Review audit trail logging only
• •validation - Review input validation only
• •errors - Review error handling only
• •uncertainty - Review uncertainty communication only
• •Or omit to review all areas

What This Skill Does

1. •Checks for proper audit trail logging
2. •Verifies input validation
3. •Reviews error handling for safety
4. •Evaluates uncertainty communication

Context

Medical imaging software has special requirements:

• •Actions must be auditable
• •Errors must not cause silent failures
• •User inputs must be validated
• •Uncertainty should be communicated clearly

While this extension is not FDA-regulated, following best practices:

• •Builds good habits for medical software development
• •Reduces risk of errors in clinical use
• •Improves overall code quality

Focus Areas

Audit Trail Logging

Check that user actions are logged:

• •Parameter changes (algorithm, threshold, etc.)
• •Painting actions (add, erase)
• •Configuration changes
• •Error conditions

Input Validation

Check parameter validation:

• •Range checking (min/max values)
• •Type checking (int vs float)
• •Bounds checking (brush radius > 0)
• •Reasonable defaults

Error Handling for Safety

Check error handling:

• •No silent failures
• •Errors reported to user
• •Graceful degradation
• •Recovery options

Uncertainty Communication

Check uncertainty handling:

• •Algorithm confidence not overstated
• •Limitations documented
• •User warned of edge cases
• •Results marked as suggestions, not diagnoses

Execution Steps

Step 1: Audit Logging Analysis

Search for logging in user action handlers:

# Find event handlers
grep -rn "def.*Event\|def on\|def handle" --include="*.py"

# Check for logging in those functions
# Each should have logging.info/debug calls

Step 2: Input Validation Analysis

Search for parameter setting and validation:

# Find parameter setters
grep -rn "setParameter\|self\.\w\+ =" --include="*.py"

# Check for validation
grep -rn "if.*<\|if.*>\|raise ValueError\|raise TypeError" --include="*.py"

Step 3: Error Handling Analysis

Check error handling patterns:

# Find exception handlers
grep -rn "except\|try:" --include="*.py"

# Check for user notification
grep -rn "slicer.util.errorDisplay\|QMessageBox\|logging.error" --include="*.py"

Step 4: Uncertainty Analysis

Review user-facing text:

• •Help text
• •Status messages
• •Result descriptions

Step 5: Generate Report

Create report in reviews/reports/<timestamp>_medical-compliance/:

• •report.json - Machine-readable findings
• •report.md - Human-readable summary

Compliance Categories

Severity for Medical Context

Output Example

## Medical Compliance Review

**Date:** 2026-01-26T14:30:00
**Focus:** Medical imaging software best practices

### Summary
- Critical: 0
- High: 3
- Medium: 8
- Low: 5

### Audit Trail Logging

#### AUDIT_MISSING [high] SegmentEditorEffect.py:1250
- **Action:** Algorithm change by user
- **Issue:** No logging when user changes algorithm selection
- **Impact:** Cannot audit what algorithm was used
- **Suggestion:** Add `logging.info(f"Algorithm changed to {algorithm}")`

#### AUDIT_MISSING [medium] SegmentEditorEffect.py:890
- **Action:** Brush size change
- **Issue:** Brush size changes not logged
- **Suggestion:** Log parameter changes at debug level

### Input Validation

#### VALIDATION_MISSING [high] SegmentEditorEffect.py:1500
- **Parameter:** sigma value for level set
- **Issue:** No range validation, could cause algorithm failure
- **Suggestion:** Add validation: `if sigma <= 0: raise ValueError("sigma must be positive")`

### Error Handling

#### SILENT_FAILURE [high] SegmentEditorEffect.py:3501
- **Context:** Generic exception in algorithm
- **Issue:** Error is logged but processing continues silently
- **Impact:** User may not know segmentation failed
- **Suggestion:** Show user error dialog or clear visual indication

### Uncertainty Communication

#### UNCERTAINTY_UNCLEAR [medium] Help text
- **Text:** "Segments based on image intensity"
- **Issue:** No mention of limitations or typical accuracy
- **Suggestion:** Add note about algorithm suitability for different tissue types

Follow-up Actions

Based on the review, you may want to:

• •Add logging to user action handlers
• •Add parameter validation
• •Improve error notifications
• •Add uncertainty notes to help text