mejbaurbahar/awesome-qa-agent-skills/42crunch-api-security-testing

42Crunch API Security Testing

Overview

42Crunch provides automated, contract-driven API security testing. It enforces security at the OpenAPI/Swagger definition layer and verifies live implementation conformance, enabling an automated Audit -> Scan -> Remediate -> Validate cycle directly within AI-assisted developer workflows.

The Continuous API Security Loop

      ┌───────────────┐
      │  1. AUDIT     │  Static contract analysis of OpenAPI / Swagger specs
      └───────┬───────┘
              ▼
      ┌───────────────┐
      │  2. SCAN      │  Dynamic conformance scanning of active endpoints
      └───────┬───────┘
              ▼
      ┌───────────────┐
      │ 3. REMEDIATE  │  Propose exact contract & code fixes (BOLA, schemas)
      └───────┬───────┘
              ▼
      ┌───────────────┐
      │ 4. VALIDATE   │  Re-audit to verify score meets threshold (>85/100)
      └───────────────┘

1. OpenAPI Security Audit Criteria (Scoring 0–100)

A high 42Crunch audit score requires comprehensive, strict contract definitions:

Authentication & Authorization

• Every endpoint (except explicit public endpoints like login/register) must define a security requirement.
• Global security definitions must declare granular scopes (e.g. read:users, write:orders).
• Unauthenticated endpoints must explicitly set security: [] with documented rationale.

Strict Schema Definition & Input Constraints

• Strings: Must specify maxLength and where applicable a restrictive pattern (regex) or format (e.g., uuid, email, uri).
• Numbers/Integers: Must specify minimum and maximum boundaries to prevent integer overflows or resource exhaustion.
• Arrays: Must specify maxItems and minItems to prevent unbounded payload attacks (DoS).
• Objects: Must specify additionalProperties: false to mitigate Mass Assignment vulnerabilities.
• Enums: Restrict input values to known permitted sets rather than arbitrary strings.

Response Contracts

• Explicit definitions for all HTTP response codes:
• 200 / 201 / 204 with strict output schemas.
• 400 Bad Request with standardized validation error schema.
• 401 Unauthorized & 403 Forbidden for auth failures.
• 404 Not Found for resource lookups.
• 429 Too Many Requests with rate limit headers.
• 500 Internal Server Error with generic, non-leaking error responses (never stack traces).

2. OWASP API Security Top 10 (2023) Mapping

3. Remediation Patterns (Before vs. After)

Example 1: Preventing Mass Assignment & Unbounded Strings

❌ Vulnerable Spec (Audit Score: ~35/100)

paths:
  /users:
    post:
      summary: Create user
      requestBody:
        content:
          application/json:
            schema:
              type: object
              properties:
                username:
                  type: string
                email:
                  type: string
                role:
                  type: string

✅ Remediated Spec (Audit Score: ~95/100)

paths:
  /users:
    post:
      summary: Create user
      security:
        - BearerAuth: ["users:create"]
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              additionalProperties: false
              required:
                - username
                - email
              properties:
                username:
                  type: string
                  minLength: 3
                  maxLength: 30
                  pattern: '^[a-zA-Z0-9_-]+$'
                email:
                  type: string
                  format: email
                  maxLength: 255
      responses:
        '201':
          description: User created
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/UserResponse'
        '400':
          $ref: '#/components/responses/400BadRequest'
        '401':
          $ref: '#/components/responses/401Unauthorized'

4. Automation & CI/CD Shift-Left

GitHub Actions Integration

name: 42Crunch API Security Audit
 
on:
  push:
    paths:
      - '**/openapi.yaml'
      - '**/openapi.json'
  pull_request:
 
jobs:
  audit:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: 42Crunch API Security Audit
        uses: 42Crunch/api-security-audit-action@v3
        with:
          api-token: ${{ secrets.SECURITY_API_TOKEN }}
          min-score: 85

CLI Quick Audit

When 42Crunch CLI (42c) or VS Code extension is installed:

# Run local contract audit
42c audit openapi.yaml --min-score 85
 
# Conformance scan against live test server
42c scan openapi.yaml --target http://localhost:8080

5. Review Checklist for Agents

When reviewing or authoring an API contract:

1. [ ] Is security defined at root and per-endpoint?
2. [ ] Does every string have maxLength and a pattern or format?
3. [ ] Does every numeric value have minimum and maximum?
4. [ ] Does every array have maxItems?
5. [ ] Is additionalProperties: false set on all request payload schemas?
6. [ ] Are all error status codes (400, 401, 403, 404, 429, 500) documented with schemas?
7. [ ] Are internal/admin paths guarded by distinct role scopes?
8. [ ] Has the audit score achieved ≥ 85/100?