application-inspector

name	application-inspector
description	Run Microsoft Application Inspector for technology profiling and security feature detection. Use when analyzing technology stack, finding crypto/auth patterns, detecting sensitive API usage, or creating security posture reports.
allowed-tools	["Bash","Read","Glob","Grep"]

Microsoft Application Inspector

When to Use Application Inspector

Ideal scenarios:

• Technology stack profiling and discovery
• Identifying security-relevant code patterns (authentication, cryptography, logging)
• Pre-audit reconnaissance to understand codebase capabilities
• Detecting use of sensitive APIs and security controls
• Creating software composition reports
• Finding data handling patterns (PII, credentials, encryption)

Complements other tools:

• Use before Semgrep/CodeQL for context on what security patterns exist
• Combine with SARIF Issue Reporter for detailed findings analysis
• Use alongside SCA tools to understand both dependencies and implementation

When NOT to Use

Do NOT use this skill for:

• Finding specific vulnerabilities (use Semgrep, CodeQL, or SAST tools)
• Dependency vulnerability scanning (use OSV-Scanner or Depscan)
• Secrets detection (use Gitleaks)
• IaC security analysis (use KICS)
• Deep data flow analysis

Installation

# .NET tool (recommended)
dotnet tool install --global Microsoft.CST.ApplicationInspector.CLI

# Update
dotnet tool update --global Microsoft.CST.ApplicationInspector.CLI

# Verify
appinspector --version

# Docker
docker pull mcr.microsoft.com/app-inspector
docker run -v ${PWD}:/app mcr.microsoft.com/app-inspector analyze -s /app -f sarif -o /app/results.sarif

Core Workflow

1. Quick Analysis

# Analyze directory with default rules
appinspector analyze -s /path/to/code -f html -o report.html

# Text summary
appinspector analyze -s /path/to/code -f text

# JSON output
appinspector analyze -s /path/to/code -f json -o results.json

2. SARIF Output

# Generate SARIF report
appinspector analyze -s /path/to/code \
  --output-file-format sarif \
  --output-file-path results.sarif

# With custom rules
appinspector analyze -s /path/to/code \
  -r /path/to/custom-rules \
  --output-file-format sarif \
  --output-file-path results.sarif

# Single-threaded for stability
appinspector analyze -s /path/to/code \
  --single-threaded \
  --file-timeout 500000 \
  --output-file-format sarif \
  --output-file-path results.sarif

3. Security-Focused Analysis

# Focus on security features
appinspector analyze -s /path/to/code \
  -t "Authentication,Cryptography,Authorization" \
  -f json -o security-features.json

# Exclude test files
appinspector analyze -s /path/to/code \
  -e "test,tests,spec,__pycache__" \
  -f sarif -o results.sarif

Understanding Output

Tag Categories

Application Inspector detects patterns across categories:

Category	Examples
Authentication	OAuth, JWT, Session management, Password handling
Cryptography	AES, RSA, Hashing, Key derivation, Random generation
Authorization	RBAC, ACL, Permission checks, Policy enforcement
Data.PII	Email, SSN, Credit card, Phone numbers
Data.Credentials	API keys, Passwords, Tokens, Certificates
CloudServices	AWS, Azure, GCP API usage
Framework	Express, Django, Spring, ASP.NET
Database	SQL, NoSQL, ORM usage

Severity Levels

• Critical: High-risk patterns (hardcoded secrets, weak crypto)
• Important: Security-relevant code requiring review
• Moderate: Potentially sensitive functionality
• ManualReview: Patterns requiring human analysis
• BestPractice: Recommended patterns found

Custom Rules

Rule Structure

{
  "name": "Detect hardcoded API keys",
  "id": "DS123456",
  "description": "Identifies potential hardcoded API keys",
  "tags": [
    "Data.Credentials.APIKey"
  ],
  "severity": "Critical",
  "patterns": [
    {
      "pattern": "api[_-]?key\\s*=\\s*['\"][a-zA-Z0-9]{20,}['\"]",
      "type": "regex",
      "confidence": "High",
      "scopes": [
        "code"
      ]
    }
  ]
}

Rule File Format

Create custom-rules.json:

[
  {
    "name": "AWS Access Key",
    "id": "DS001",
    "tags": ["Data.Credentials.AWS"],
    "severity": "Critical",
    "patterns": [
      {
        "pattern": "AKIA[0-9A-Z]{16}",
        "type": "regex",
        "confidence": "High"
      }
    ]
  }
]

Use with:

appinspector analyze -s /code -r custom-rules.json -f sarif -o results.sarif

Verification Commands

# Verify rules
appinspector verify-rules -r /path/to/rules

# Test specific rule
appinspector verify-rules -r custom-rules.json

# List default rules
appinspector exportrules -o default-rules.json

CI/CD Integration (GitHub Actions)

name: Application Inspector

on:
  push:
    branches: [main]
  pull_request:
  schedule:
    - cron: '0 0 1 * *'

jobs:
  analyze:
    runs-on: ubuntu-latest

    steps:
      - uses: actions/checkout@v4

      - name: Setup .NET
        uses: actions/setup-dotnet@v4
        with:
          dotnet-version: '8.0'

      - name: Install Application Inspector
        run: dotnet tool install --global Microsoft.CST.ApplicationInspector.CLI

      - name: Run Analysis
        run: |
          appinspector analyze \
            -s ${{ github.workspace }} \
            --output-file-format sarif \
            --output-file-path results.sarif \
            --single-threaded \
            --disable-archive-crawling

      - name: Upload SARIF
        uses: github/codeql-action/upload-sarif@v3
        with:
          sarif_file: results.sarif
          category: application-inspector

      - name: Upload Results
        uses: actions/upload-artifact@v4
        if: always()
        with:
          name: application-inspector-results
          path: results.sarif

Configuration

Performance Options

# Single-threaded (more stable)
appinspector analyze -s /code --single-threaded

# Increase timeout for large files
appinspector analyze -s /code --file-timeout 500000

# Disable archive scanning (faster)
appinspector analyze -s /code --disable-archive-crawling

# Process only specific languages
appinspector analyze -s /code -l "javascript,typescript,python"

Exclusions

# Exclude paths
appinspector analyze -s /code \
  -e "node_modules,vendor,dist,build,__pycache__"

# Exclude file patterns
appinspector analyze -s /code \
  -e "*.min.js,*.test.js,*.spec.ts"

Common Use Cases

1. Pre-Audit Technology Discovery

# Generate comprehensive technology report
appinspector analyze -s /code -f html -o tech-report.html

# Review report to understand:
# - What frameworks are used
# - What crypto libraries are present
# - How authentication is implemented
# - What cloud services are integrated

2. Security Feature Inventory

# Find all security-relevant patterns
appinspector analyze -s /code \
  -t "Authentication,Authorization,Cryptography,Data.Credentials" \
  -f json -o security-inventory.json

3. Compliance Scanning

# Detect PII handling
appinspector analyze -s /code \
  -t "Data.PII" \
  -f sarif -o pii-report.sarif

# Find credential usage
appinspector analyze -s /code \
  -t "Data.Credentials" \
  -f sarif -o credentials-report.sarif

Interpreting Results

SARIF Structure

Application Inspector SARIF includes:

• Rules: Each detected pattern/tag
• Results: Specific code locations matching patterns
• Properties: Confidence, severity, tags
• Locations: File path, line number, code snippet

Filter by Severity

# Use SARIF tools to filter
pip install sarif-tools

# Extract critical findings only
sarif summary results.sarif --level error

# Filter by tag
sarif filter --level error --rule-id "DS.*Credentials.*" results.sarif

Limitations

• Not a vulnerability scanner: Identifies patterns, not exploits
• False positives: Regex-based detection can flag legitimate code
• Performance: Large codebases may require single-threaded mode
• Language coverage: Better for common languages (JS, Python, C#, Java)
• No data flow: Can't track how data moves through application

Rationalizations to Reject

Shortcut	Why It's Wrong
"AppInspector found crypto, so it's secure"	Finding crypto usage doesn't mean it's implemented correctly; manual review required
"No credentials found = code is clean"	Pattern-based detection misses obfuscated or dynamically constructed secrets
"High confidence = definite issue"	High confidence means pattern match strength, not security impact
"Skip single-threaded mode for speed"	Multi-threaded can crash on complex codebases; stability > speed
"HTML report is enough"	SARIF output enables integration with other tools and automated workflows

References

• Repository: https://github.com/microsoft/ApplicationInspector
• Documentation: https://github.com/microsoft/ApplicationInspector/wiki
• Default Rules: https://github.com/microsoft/ApplicationInspector/tree/main/AppInspector/rules
• SARIF Spec: https://docs.oasis-open.org/sarif/sarif/v2.1.0/sarif-v2.1.0.html