> ## Documentation Index
> Fetch the complete documentation index at: https://docs.codethreat.com/llms.txt
> Use this file to discover all available pages before exploring further.

# On-Demand Security Analysis

> Request AI-powered security reviews for specific code or scenarios

Beyond automatic scanning, request AI-powered security analysis for specific files, features, or scenarios.

## Agentic Code Analysis (AGENTIC\_SAST)

Beyond traditional SAST rules, CodeThreat's AI can analyze code with human-like intelligence.

**What it finds that traditional SAST can't**:

* Authorization flaws (business logic vulnerabilities)
* Race conditions in concurrent code
* Complex authentication bypass vectors
* Business logic vulnerabilities
* Context-dependent security issues

### When to Use Agentic Analysis

<CardGroup cols={2}>
  <Card title="Critical Features" icon="magnifying-glass">
    Deep analysis of authentication, payment processing, admin functions
  </Card>

  <Card title="Pre-Release Audits" icon="shield-check">
    Comprehensive review before deploying to production
  </Card>

  <Card title="Incident Investigation" icon="bug">
    Understand how a vulnerability was introduced
  </Card>

  <Card title="Fix Verification" icon="circle-check">
    Verify security fixes actually eliminate the vulnerability
  </Card>
</CardGroup>

***

## Request Analysis

<Steps>
  <Step title="Navigate to Repository">
    Open the repository you want to analyze
  </Step>

  <Step title="Click Security Analysis">
    Click **AI Security Analysis** in the repository menu
  </Step>

  <Step title="Choose Scope">
    Select what to analyze:

    * Specific file
    * Directory
    * Feature (related files)
    * Full repository
  </Step>

  <Step title="Specify Focus (Optional)">
    Provide context like:

    * "Analyze authentication logic for bypass vulnerabilities"
    * "Review payment processing for security issues"
    * "Check if admin endpoints are properly protected"
  </Step>

  <Step title="Submit Request">
    Click **Request Analysis**
  </Step>

  <Step title="Review Results">
    Results appear in **AI Analysis Reports** tab, typically within 5-15 minutes
  </Step>
</Steps>

***

## Analysis Report

AI-generated security analysis report includes:

### Executive Summary

* Overall security posture
* Critical, High, Medium, Low issue counts
* Key findings summary
* Analysis duration

### Detailed Findings

Each finding includes:

* **Vulnerability description**: Clear explanation
* **Location**: File, line number, code snippet
* **Attack scenario**: How an attacker could exploit this
* **Impact**: What data/functionality is at risk
* **Recommendation**: Specific steps to fix
* **Code example**: Secure implementation
* **Priority**: Critical, High, Medium, Low
* **Confidence**: AI's certainty level

### Architecture Insights

* Security strengths
* Security weaknesses
* Recommendations for improvement

***

## Best Practices

* Use for critical features before release
* Request analysis after major refactoring
* Analyze authentication and authorization logic
* Review payment processing code
* Verify security fixes are effective

***

## Next Steps

<CardGroup cols={2}>
  <Card title="AI Overview" icon="brain" href="/ai/overview">
    Learn about AI-powered analysis
  </Card>

  <Card title="PR Reviews" icon="code-pull-request" href="/ai/pr-reviews">
    AI-powered PR reviews
  </Card>
</CardGroup>