Numeric Patterns

Advanced numeric comparison and validation patterns.

MCP Aegis provides comprehensive numeric comparison patterns for validating numeric values, counts, scores, measurements, and calculations returned by MCP servers. These patterns support both integer and floating-point numbers.

Important Note

All numeric patterns automatically parse string values to numbers. If the value cannot be parsed as a number, the pattern will return false.

Comparison Patterns

Greater Than

Use "match:greaterThan:N" to validate that a numeric value is greater than the specified threshold.

yaml

# Validate that a count is greater than expected minimum
expect:
  response:
    result:
      count: "match:greaterThan:10"        # Must be > 10
      score: "match:greaterThan:95.5"      # Must be > 95.5
      percentage: "match:greaterThan:0"     # Must be positive

Greater Than or Equal

Use "match:greaterThanOrEqual:N" to validate that a numeric value is greater than or equal to the threshold.

yaml

# Validate minimum values (inclusive)
expect:
  response:
    result:
      rating: "match:greaterThanOrEqual:1"    # Must be >= 1
      progress: "match:greaterThanOrEqual:0"   # Must be >= 0 (non-negative)
      temperature: "match:greaterThanOrEqual:-273.15"  # Above absolute zero

Less Than

Use "match:lessThan:N" to validate that a numeric value is less than the specified threshold.

yaml

# Validate maximum limits
expect:
  response:
    result:
      errorRate: "match:lessThan:0.01"      # Must be < 1%
      responseTime: "match:lessThan:1000"   # Must be < 1000ms
      usage: "match:lessThan:100"           # Must be < 100%

Less Than or Equal

Use "match:lessThanOrEqual:N" to validate that a numeric value is less than or equal to the threshold.

yaml

# Validate maximum values (inclusive)
expect:
  response:
    result:
      percentage: "match:lessThanOrEqual:100"  # Must be <= 100%
      priority: "match:lessThanOrEqual:10"     # Must be <= 10
      fileSize: "match:lessThanOrEqual:1024"   # Must be <= 1KB

Range Patterns

Between (Inclusive Range)

Use "match:between:MIN:MAX" to validate that a numeric value falls within a specific range (inclusive).

yaml

# Validate values within acceptable ranges
expect:
  response:
    result:
      httpStatus: "match:between:200:299"     # Success status codes (200-299)
      temperature: "match:between:-10:40"     # Temperature range (-10°C to 40°C)
      rating: "match:between:1:5"             # Rating scale (1 to 5 stars)
      percentage: "match:between:0:100"       # Percentage (0% to 100%)

Range (Alias for Between)

Use "match:range:MIN:MAX" as an alias for between. Both patterns work identically.

yaml

# Same functionality as between, different syntax preference
expect:
  response:
    result:
      score: "match:range:0:100"           # Score range (0 to 100)
      port: "match:range:1024:65535"       # Valid port range
      latitude: "match:range:-90:90"       # Valid latitude range

Production Examples

Here are real-world examples adapted from MCP Aegis's test suite:

API Response Validation

yaml

# Validate API response contains numeric data within expected ranges
- it: "should validate API response metrics are within acceptable ranges"
  request:
    jsonrpc: "2.0"
    id: "api-metrics"
    method: "tools/call"
    params:
      name: "get_api_metrics"
      arguments: {}
  expect:
    response:
      jsonrpc: "2.0"
      id: "api-metrics"
      result:
        responseTime: "match:lessThan:500"        # Must be < 500ms
        successRate: "match:greaterThan:99"       # Must be > 99%
        errorCount: "match:lessThanOrEqual:10"    # Must be <= 10 errors
        uptime: "match:between:99:100"            # 99-100% uptime
        requestCount: "match:greaterThanOrEqual:0" # Non-negative
        averageLatency: "match:range:0:1000"      # 0-1000ms range

File System Validation

yaml

# Example from filesystem server testing numeric file properties
- it: "should validate file statistics are within reasonable bounds"
  request:
    jsonrpc: "2.0"
    id: "file-stats"
    method: "tools/call"
    params:
      name: "get_file_stats"
      arguments:
        path: "./large-file.txt"
  expect:
    response:
      jsonrpc: "2.0"
      id: "file-stats"
      result:
        sizeBytes: "match:greaterThan:0"          # File must have content
        maxSizeBytes: "match:lessThan:10485760"   # Must be < 10MB
        lineCount: "match:between:1:1000"        # Reasonable line count
        modifiedTimestamp: "match:greaterThan:0"  # Valid timestamp

Mathematical Calculation Validation

yaml

# Example from multi-tool server calculator testing
- it: "should validate calculation results are mathematically correct"
  request:
    jsonrpc: "2.0"
    id: "calc-validation"
    method: "tools/call"
    params:
      name: "calculator"
      arguments:
        operation: "divide"
        a: 100
        b: 3
  expect:
    response:
      jsonrpc: "2.0"
      id: "calc-validation"
      result:
        result: "match:between:33:34"             # 100/3 ≈ 33.33
        precision: "match:greaterThan:2"          # At least 2 decimal places
        isExact: "match:type:boolean"             # Type validation
        # Combining numeric with other patterns:
        formatted: "match:contains:33.33"         # String representation

Combining with Pattern Negation

All numeric patterns work with pattern negation using "match:not:":

yaml

expect:
  response:
    result:
      errorRate: "match:not:greaterThan:5"      # NOT > 5 (i.e., <= 5)
      status: "match:not:between:400:599"       # NOT 4xx-5xx (i.e., success)
      timeout: "match:not:lessThan:1000"        # NOT < 1000 (i.e., >= 1000)
      score: "match:not:range:0:50"             # NOT 0-50 (i.e., > 50)

Error Handling

Numeric patterns handle edge cases gracefully:

Non-numeric strings: Return false
Null/undefined values: Return false
Invalid range syntax: Return false
String numbers: Automatically parsed (e.g., "42" becomes 42)

yaml

# These will all evaluate to false:
invalidValue: "match:greaterThan:10"        # When actual is "abc"
nullValue: "match:between:1:10"             # When actual is null
malformedRange: "match:between:invalid"     # Invalid range format

Common Use Cases

Performance Testing: Response times, throughput, memory usage
API Validation: HTTP status codes, pagination counts, rate limits
Data Quality: Score ranges, percentages, measurement bounds
Business Logic: Price ranges, quantity limits, rating scales
System Metrics: CPU usage, disk space, error rates
Mathematical Operations: Calculation results, statistical measures

Best Practice

Combine numeric patterns with other pattern types for comprehensive validation. For example, use type validation first, then numeric ranges for robust testing.