Handling Sensitive Data in Code - Secrets Management | OneNoughtOne

0/246

System Design (LLD)Secrets Management

Secrets Management

LevelIntermediate

Duration60 mins

TopicSecrets Management

1 / 4

Handling Sensitive Data in Code

The Silent Breach Waiting to Happen

In 2019, a major financial services company suffered a catastrophic data breach—not through sophisticated hacking, not through zero-day exploits, but through a single database password committed to a public GitHub repository. The breach exposed 106 million customer records and cost the company over $80 million in fines alone.

This scenario repeats itself with alarming regularity across the technology industry. According to GitGuardian's State of Secrets Sprawl report, over 10 million secrets were detected in public GitHub commits in a single year—an increase of 67% from the previous year. API keys, database credentials, private certificates, and authentication tokens flow freely through version control systems, log files, error messages, and configuration dumps.

The uncomfortable truth is that most security breaches involving secrets are entirely preventable. They don't require advanced security expertise or expensive tooling. They require discipline, awareness, and proper design patterns—precisely what this module will teach you.

What You Will Learn

By the end of this page, you will understand: what constitutes sensitive data in software systems, why secrets exposure is so dangerous and common, the fundamental principles of secrets handling, threat modeling for secrets, and immediate practical steps to audit and protect sensitive data in your codebase. This knowledge forms the foundation for the entire secrets management discipline.

Defining Sensitive Data

Before we can protect sensitive data, we must precisely define what constitutes a "secret" in the context of software systems. The definition is broader than most developers initially assume, and failure to recognize all forms of sensitive data is often the root cause of leaks.

The Fundamental Definition:

A secret is any piece of information that:

Grants access to protected resources or systems
Could enable impersonation of users, services, or systems
Would cause harm, liability, or competitive disadvantage if exposed
Is subject to regulatory protection requirements

This definition encompasses far more than passwords and API keys. Let's explore the complete taxonomy of sensitive data in modern software systems.

Comprehensive Taxonomy of Sensitive Data
Category	Examples	Exposure Risk Level	Common Leak Vectors
Authentication Credentials	Passwords, API keys, access tokens, OAuth secrets, JWT signing keys	Critical	Source code, logs, error messages, config files
Cryptographic Material	Private keys, certificates, encryption keys, HMAC secrets, key derivation salts	Critical	Key stores, backup files, deployment scripts
Infrastructure Secrets	Database connection strings, service account credentials, cloud provider keys	Critical	Environment files, CI/CD configs, container images
Personal Identifiable Information (PII)	SSNs, credit card numbers, medical records, biometric data	High (Regulatory)	Database dumps, logs, cache systems
Business Confidential Data	Trade secrets, pricing algorithms, customer lists, strategic plans	High (Business)	Application logs, analytics pipelines, debug outputs
Session & Authorization Data	Session tokens, cookies, refresh tokens, authorization grants	High	Browser storage, logs, URL parameters
Internal System Details	Internal URLs, network topology, system architecture details	Medium	Error messages, documentation, config files

The Hidden Secrets Problem

Many developers focus only on obvious secrets like database passwords while ignoring equally dangerous data. A leaked internal API endpoint might seem harmless, but combined with other information it can enable reconnaissance attacks. A leaked internal user ID format might allow enumeration attacks. Treat all internal system details as potentially sensitive.

Sensitivity Classification Framework:

Professional organizations implement formal classification systems to ensure consistent handling of sensitive data. A practical four-tier classification system works as follows:

Data Sensitivity Classifications

•PUBLIC — Information intended for public consumption. Marketing materials, public documentation, non-sensitive configuration. Can be freely logged and transmitted.
•INTERNAL — Information for internal use only. Internal documentation, development URLs, non-production data. Should not be exposed externally but can be logged internally with care.
•CONFIDENTIAL — Sensitive business or operational data. Production credentials, customer data, business logic details. Must be encrypted at rest and in transit, never logged in plaintext.
•RESTRICTED — Highest sensitivity data. Cryptographic keys, root credentials, regulated PII (health, financial). Requires hardware security modules (HSM), strict access controls, comprehensive auditing.

Why Secrets Exposure Is So Dangerous

Understanding the severity of secrets exposure requires examining the attack chains that become possible once a secret is leaked. Unlike many security vulnerabilities that require exploitation skills, exposed secrets typically enable immediate, direct access to protected resources.

The Asymmetry of Secrets Security:

Secrets security is fundamentally asymmetric:

Defense requires perfection: A single leak at any point in the software lifecycle can compromise security
Attack requires only discovery: Finding one exposed secret in millions of files grants access
Damage is often silent: Attackers use stolen credentials to blend in as legitimate users
Recovery is expensive: Rotating secrets requires updating all dependent systems simultaneously

Attack Chains Enabled by Leaked Secrets

•Lateral Movement — A leaked database credential provides access to data, which often contains additional credentials for other systems. Attackers 'pivot' from one compromised system to another, escalating access.
•Data Exfiltration — With valid credentials, attackers can extract data at leisure, often over extended periods without detection. The average breach goes undetected for 287 days.
•Service Impersonation — Leaked API keys allow attackers to make requests that appear legitimate, bypassing normal security monitoring and potentially manipulating data or systems.
•Cryptographic Bypass — Exposed encryption keys or signing secrets can render entire encryption schemes useless, allowing decryption of protected data or forgery of trusted tokens.
•Supply Chain Attacks — Cloud provider credentials or CI/CD tokens enable attackers to inject malicious code into build pipelines, affecting all downstream consumers.
•Regulatory Violation — Even without active exploitation, exposure of regulated data (PII, PHI, financial data) triggers mandatory breach notifications and regulatory penalties.

The Permanence Problem

Once a secret is exposed to version control, it is effectively permanent. Even if deleted in the next commit, the secret remains in Git history forever. Automated scanners, archive services, and attackers regularly mine Git history for secrets. The only safe response is to assume the secret is compromised and rotate it immediately.

Real-World Cost Analysis:

The financial and operational impact of secrets exposure is substantial and multi-dimensional:

Cost Components of Secrets-Related Breaches
Cost Category	Description	Typical Range
Direct Financial Loss	Fraud, theft, ransom payments	$10K - $10M+
Incident Response	Investigation, forensics, remediation	$50K - $500K
Regulatory Fines	GDPR, HIPAA, PCI-DSS violations	$100K - $50M+
Legal Liability	Lawsuits, settlements, legal fees	$100K - $100M+
Business Disruption	Downtime, emergency rotations, lost productivity	$50K - $5M
Reputation Damage	Customer churn, lost business, brand impact	Immeasurable
Long-term Monitoring	Credit monitoring, ongoing security investments	$100K - $1M/year

Common Secrets Exposure Vectors

Understanding how secrets leak is essential for prevention. Secrets exposure occurs through predictable, well-documented vectors—each requiring specific countermeasures. Let's examine the primary leak vectors in detail.

Vector 1: Source Code and Version Control

The most common and dangerous exposure vector. Developers embed secrets directly in source files for convenience during development, then forget to remove them before commit.

DANGEROUS: Hardcoded secrets examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
// ❌ CATASTROPHICALLY WRONG: Hardcoded secrets
class DatabaseService {
    // This secret is now in Git history FOREVER
    private readonly connectionString = "postgresql://admin:SuperSecretP@[email protected]:5432/customers";
    
    // API keys visible to anyone with repository access
    private readonly stripeApiKey = "sk_live_abcd1234efgh5678ijkl9012mnop";
    
    // AWS credentials - grants full cloud access
    private readonly awsAccessKey = "AKIAIOSFODNN7EXAMPLE";
    private readonly awsSecretKey = "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY";
    
    async connect(): Promise<void> {
        // Even if you delete these later, they're in Git history
        await this.pool.connect(this.connectionString);
    }
}
 
// ❌ ALSO WRONG: Secrets in configuration objects
const config = {
    jwt: {
        // Anyone who clones this repo can forge tokens
        secret: "my-super-secret-jwt-signing-key-2024",
        expiresIn: "24h"
    },
    encryption: {
        // Encryption is useless if the key is public
        key: "aes-256-encryption-key-12345678901234567890",
        algorithm: "aes-256-gcm"
    }
};
 
// ❌ WRONG: "Temporary" secrets that become permanent
// TODO: Move to environment variables before production
const TEMP_API_KEY = "api_key_12345"; // Added 3 years ago...

Vector 2: Configuration Files and Environment

Configuration files are the primary target for secrets scanning because they're designed to hold variable values. Even when developers avoid hardcoding in source, they often commit configuration files with secrets.

DANGEROUS: Configuration file secrets

# ❌ NEVER COMMIT .env files with real secrets!
 
# Database (if this file is committed, your database is compromised)
DATABASE_URL=postgresql://admin:[email protected]:5432/production
REDIS_URL=redis://:[email protected]:6379
 
# API Keys (these grant access to external services)
STRIPE_SECRET_KEY=sk_live_real_key_here
SENDGRID_API_KEY=SG.real_key_here
TWILIO_AUTH_TOKEN=real_twilio_token
 
# AWS Credentials (full cloud access)
AWS_ACCESS_KEY_ID=AKIAIOSFODNN7EXAMPLE
AWS_SECRET_ACCESS_KEY=wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
 
# JWT (anyone with this can forge authentication tokens)
JWT_SECRET=production-jwt-secret-change-me
 
# Encryption (all encrypted data can be decrypted)
ENCRYPTION_KEY=32-character-encryption-key-here

Vector 3: Logging and Error Messages

Logging is essential for debugging and monitoring, but careless logging is a major secrets exposure vector. Secrets end up in logs through error dumps, request logging, and debug statements.

Dangerous Logging Patterns

•Logging full HTTP request bodies containing auth headers
•Exception messages that include connection strings
•Debug logging of configuration at startup
•Logging user input that may contain passwords
•Stack traces that reveal secret values
•Logging entire objects without filtering sensitive fields

Safe Logging Practices

•Explicit allowlists of fields that can be logged
•Automatic redaction of known sensitive field names
•Structured logging with sensitive field markers
•Request logging that excludes Authorization headers
•Custom exception types that hide internal details
•Serialization interceptors that mask sensitive data

Defense in Depth for Logs

Implement multiple layers of log protection: 1) Don't put secrets in loggable structures in the first place, 2) Filter known sensitive patterns at the logging framework level, 3) Encrypt or restrict access to log storage, 4) Implement log retention policies that limit exposure window.

Fundamental Principles of Secrets Handling

Effective secrets management is built on a foundation of core principles. These principles guide every decision about how secrets are stored, transmitted, used, and retired. Internalize these principles, and secure secrets handling becomes intuitive rather than a compliance checklist.

The Seven Principles of Secrets Handling

•Principle 1: Separation — Secrets must be separated from code. Code is version-controlled, shared, and often public. Secrets are none of these things. They exist in a separate system with different access controls and lifecycle management.
•Principle 2: Least Privilege — Every component should have access only to the secrets it absolutely needs. A web frontend doesn't need database credentials. A logging service doesn't need payment API keys. Minimize blast radius by minimizing access.
•Principle 3: Defense in Depth — No single control should be your only protection. Secrets should be encrypted at rest AND in transit AND access-controlled AND audited AND rotated. Each layer catches what others miss.
•Principle 4: Auditability — Every secrets access should be logged and auditable. When a breach occurs, you must be able to determine which secrets were accessed, by whom, and when. Audit trails also deter insider threats.
•Principle 5: Rotation Capability — Secrets must be designed for rotation from day one. If rotating a secret requires system-wide changes, you've designed incorrectly. Secret rotation should be a routine, automated operation.
•Principle 6: Secure Defaults — Systems should fail closed, not open. If a secret is unavailable, the operation should fail rather than proceeding insecurely. Never fall back to hardcoded defaults in production.
•Principle 7: Immutable History — Assume anything that enters version control is permanent and public. Design accordingly. Never rely on 'deleting' a committed secret—rotate it instead.

Applying the Principles - A Practical Framework:

These principles translate into concrete practices for everyday development:

secrets-handling-framework.ts
TypeScript
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
/**
 * Secrets Handling Framework
 * 
 * Demonstrates the core patterns for proper secrets management
 * following all seven principles.
 */
 
// ============================================
// Principle 1: SEPARATION - Secrets come from external sources
// ============================================
 
interface SecretsProvider {
    /**
     * Retrieves a secret by its logical name.
     * Implementation is decoupled from usage.
     */
    getSecret(name: string): Promise<string>;
    
    /**
     * Checks if a secret exists without retrieving it.
     * Useful for validation without exposure.
     */
    hasSecret(name: string): Promise<boolean>;
}
 
// Different environments use different providers
// Code doesn't know or care where secrets come from
class EnvironmentSecretsProvider implements SecretsProvider {
    async getSecret(name: string): Promise<string> {
        const value = process.env[name];
        if (!value) {
            throw new SecretNotFoundError(name);
        }
        return value;
    }
    
    async hasSecret(name: string): Promise<boolean> {
        return process.env[name] !== undefined;
    }
}
 
class VaultSecretsProvider implements SecretsProvider {
    constructor(private vault: VaultClient) {}
    
    async getSecret(name: string): Promise<string> {
        // Vault handles encryption, access control, auditing
        return this.vault.read(`secret/data/${name}`);
    }
    
    async hasSecret(name: string): Promise<boolean> {
        try {
            await this.vault.read(`secret/data/${name}`);
            return true;
        } catch {
            return false;
        }
    }
}
 
// ============================================
// Principle 2: LEAST PRIVILEGE - Each service gets only what it needs
// ============================================
 
interface ServiceSecrets {
    // Only declare secrets this specific service needs
    readonly databaseUrl: string;
    readonly jwtSecret: string;
}
 
interface PaymentServiceSecrets extends ServiceSecrets {
    // Payment service also needs payment provider credentials
    readonly stripeSecretKey: string;
}
 
// Factory creates secrets objects with only required credentials
class SecretsFactory {
    constructor(private provider: SecretsProvider) {}
    
    async createForWebService(): Promise<ServiceSecrets> {
        return {
            databaseUrl: await this.provider.getSecret('DATABASE_URL'),
            jwtSecret: await this.provider.getSecret('JWT_SECRET'),
            // Note: No payment keys here - web service doesn't process payments
        };
    }
    
    async createForPaymentService(): Promise<PaymentServiceSecrets> {
        return {
            databaseUrl: await this.provider.getSecret('DATABASE_URL'),
            jwtSecret: await this.provider.getSecret('JWT_SECRET'),
            stripeSecretKey: await this.provider.getSecret('STRIPE_SECRET_KEY'),
        };
    }
}
 
// ============================================
// Principle 3: DEFENSE IN DEPTH - Multiple layers of protection
// ============================================
 
class SecureSecret {
    private value: string;
    private accessLog: AccessLogEntry[] = [];
    
    constructor(value: string) {
        // Store encrypted in memory (not foolproof, but adds layer)
        this.value = this.encrypt(value);
    }
    
    /**
     * Accessing the secret is explicit and audited
     */
    expose(purpose: string, accessor: string): string {
        // Log every access
        this.accessLog.push({
            accessor,
            purpose,
            timestamp: new Date(),
        });
        
        // Return decrypted value
        return this.decrypt(this.value);
    }
    
    /**
     * Never accidentally leak secrets through logging or serialization
     */
    toString(): string {
        return '[REDACTED]';
    }
    
    toJSON(): string {
        return '[REDACTED]';
    }
    
    // For debugging without exposing
    get length(): number {
        return this.decrypt(this.value).length;
    }
    
    private encrypt(value: string): string {
        // Simplified - use proper encryption in production
        return Buffer.from(value).toString('base64');
    }
    
    private decrypt(encrypted: string): string {
        return Buffer.from(encrypted, 'base64').toString();
    }
}
 
// ============================================
// Principle 6: SECURE DEFAULTS - Fail closed
// ============================================
 
class SecretNotFoundError extends Error {
    constructor(secretName: string) {
        // Never reveal the secret name in production errors
        super(`Required configuration not found: ${
            process.env.NODE_ENV === 'development' ? secretName : '[REDACTED]'
        }`);
        this.name = 'SecretNotFoundError';
    }
}
 
async function initializeApplication(
    secretsProvider: SecretsProvider
): Promise<void> {
    // Validate ALL required secrets before starting
    const required = ['DATABASE_URL', 'JWT_SECRET', 'ENCRYPTION_KEY'];
    
    for (const secret of required) {
        if (!(await secretsProvider.hasSecret(secret))) {
            // Application fails to start - secure default
            throw new SecretNotFoundError(secret);
        }
    }
    
    // Only proceed if all secrets are available
    console.log('All required secrets validated. Starting application...');
}

Principles Over Tools

Tools and technologies change, but these principles remain constant. Whether you use HashiCorp Vault, AWS Secrets Manager, Azure Key Vault, or environment variables, the same principles apply. Master the principles, and you can implement secure secrets handling with any technology stack.

Threat Modeling for Secrets

Effective secrets protection requires understanding WHO might attack, WHAT they might target, and HOW they might attempt access. Threat modeling provides a structured approach to identifying and prioritizing secrets risks.

The STRIDE Model Applied to Secrets:

STRIDE is a threat modeling framework developed by Microsoft. Let's apply each category specifically to secrets management:

STRIDE Threat Model for Secrets Management
Threat	Description	Secrets-Specific Example	Mitigation
Spoofing	Pretending to be another user/system	Using leaked credentials to impersonate authorized services	MFA, short-lived tokens, certificate-based authentication
Tampering	Modifying data without authorization	Changing secret values in transit to inject malicious credentials	Encryption in transit (TLS), message signing, integrity verification
Repudiation	Denying actions without proof	Claiming you didn't access a secret when you did	Comprehensive audit logging, immutable audit trails
Information Disclosure	Exposing information to unauthorized parties	Secrets appearing in logs, error messages, or version control	Encryption at rest, access controls, redaction, secrets scanning
Denial of Service	Making resources unavailable	Overwhelming secrets management infrastructure to prevent legitimate access	Rate limiting, redundancy, caching (where appropriate)
Elevation of Privilege	Gaining unauthorized capabilities	Using a low-privilege secret to access higher-privilege secrets	Least privilege, secret isolation, privilege separation

Attacker Profiles:

Different attackers have different capabilities and motivations. Understanding attacker profiles helps prioritize defenses:

Common Attacker Profiles

•Opportunistic Scanners — Automated bots that scan public repositories, paste sites, and leaked data for secrets patterns. Low sophistication, high volume. Defense: Don't commit secrets, secrets scanning.
•External Attackers — Attempt to breach perimeter defenses. May use phishing, exploits, or social engineering. Defense: Network security, access controls, security awareness training.
•Malicious Insiders — Current employees with legitimate access who abuse it. Most dangerous because they bypass perimeter defenses. Defense: Least privilege, auditing, anomaly detection.
•Negligent Insiders — Well-meaning employees who accidentally expose secrets through poor practices. Most common source of breaches. Defense: Training, tooling, automated guardrails.
•Compromised Accounts — Legitimate accounts taken over through credential theft. Act with the privileges of the compromised user. Defense: MFA, session management, behavioral analysis.
•Supply Chain Attackers — Compromise dependencies or build systems to inject malicious code. Highly sophisticated, high impact. Defense: Dependency scanning, build integrity, secret isolation.

The Insider Threat

Studies consistently show that insider threats—both malicious and negligent—cause more secrets exposure than external attacks. Your colleagues with legitimate access are often the biggest risk. This isn't about distrust; it's about designing systems that make mistakes hard and malice detectable.

Practical First Steps

Theory is essential, but let's translate these principles into immediate, actionable steps you can take today to improve your codebase's secrets hygiene.

Immediate Actions Checklist

•Audit Your .gitignore — Ensure .env, .env.*, *.pem, *.key, and other secrets files are ignored. Check that these lines are present and not accidentally commented out.
•Scan Git History — Run a tool like git-secrets, truffleHog, or gitleaks against your repository to find existing secrets in history. Any found secrets should be rotated immediately.
•Install Pre-commit Hooks — Set up pre-commit hooks that scan for secrets patterns before allowing commits. This prevents new secrets from entering the repository.
•Audit Environment Variable Usage — List all environment variables your application uses. Document which are secrets vs. configuration. Ensure secrets don't have default values in code.
•Review Logging Code — Search for log statements that might include sensitive data. Look for request body logging, configuration dumps, and error messages that include connection strings.
•Implement Secret Types — Create wrapper types for secrets (like the SecureSecret class above) that prevent accidental logging and serialization.
•Enable Repository Scanning — If using GitHub, enable secret scanning. If using other platforms, integrate a third-party scanning tool into your CI/CD pipeline.
•Document Secrets Inventory — Create a list of all secrets used by your system, their sources, rotation schedules, and who should have access. This is essential for incident response.

.gitignore template for secrets protection

.gitignore

# ============================================
# SECRETS PROTECTION - NEVER REMOVE THESE LINES
# ============================================
 
# Environment files (may contain secrets)
.env
.env.*
.env.local
.env.*.local
*.env
 
# Key files and certificates
*.pem
*.key
*.p12
*.pfx
*.jks
*.keystore
*.crt
*.cer
 
# AWS credentials
.aws/credentials
aws-credentials*
 
# Google Cloud
*.json  # Be careful - only ignore service account JSON
gcloud-*.json
service-account*.json
 
# Terraform state (contains secrets)
*.tfstate
*.tfstate.*
.terraform/
 
# IDE-specific files that may contain configs
.idea/
.vscode/settings.json
*.sublime-workspace
 
# Local development databases
*.sqlite
*.db
 
# Docker secrets
secrets/
.secrets/
 
# Backup files that may contain secrets
*.bak
*.backup
*.old
 
# Log files (may contain exposed secrets)
*.log
logs/
 
# ============================================
# PROJECT-SPECIFIC ADDITIONS
# Add any project-specific secret files below
# ============================================

Start Small, Be Consistent

You don't need to implement enterprise-grade secrets management immediately. Start with the basics: don't commit secrets, use environment variables, rotate compromised secrets. Build from there as your team matures and your system scales.

Summary: Handling Sensitive Data in Code

We've established the foundational knowledge for secrets management. Let's consolidate the key takeaways before moving forward:

Key Takeaways

•Secrets are broader than passwords — Authentication credentials, cryptographic keys, PII, and even internal system details require protection.
•Exposure is permanent — Once a secret enters version control, assume it's compromised forever. Rotation is the only safe response.
•Common vectors are predictable — Source code, configuration files, logs, and error messages are the primary leak channels. Design for these specific threats.
•Seven principles guide decisions — Separation, least privilege, defense in depth, auditability, rotation capability, secure defaults, and immutable history.
•Threat modeling informs prioritization — Understanding who attacks, what they target, and how they operate helps focus security investments.
•Start with fundamentals — Basic hygiene (gitignore, scanning, pre-commit hooks) prevents the vast majority of accidental exposure.

What's next:

Now that we understand what sensitive data is and why protecting it matters, we'll explore the crucial distinction between configuration and secrets. Understanding this distinction is essential for designing systems that are both flexible and secure—allowing configuration to be managed openly while secrets remain protected.

Page Complete

You now understand the comprehensive landscape of sensitive data in software systems. You can identify secrets, understand exposure risks, apply fundamental protection principles, and take immediate action to secure your codebase. Next, we'll explore the critical distinction between configuration and secrets.