Skip to content

Security: varun369/SuperLocalMemoryV2

SECURITY.md

Security Policy

SuperLocalMemory V2 Security

This document outlines the security policy for SuperLocalMemory V2, including supported versions, vulnerability reporting procedures, and our privacy-first approach.


Table of Contents


Supported Versions

Currently Supported

Version Status Support Level Security Updates
2.0.x Active Full support Yes
1.x.x Legacy Limited Critical only

Support Timeline

  • Active support: Latest 2.x release receives all security updates
  • Legacy support: Previous major version (1.x) receives critical security fixes only
  • End of life: Versions older than 1.x are no longer supported

Recommendation: Always use the latest 2.x release for best security.


Reporting Vulnerabilities

Security-First Approach

We take security seriously. If you discover a security vulnerability, please report it responsibly.

Reporting Process

DO NOT create a public GitHub issue for security vulnerabilities.

Instead, follow these steps:

Step 1: Private Disclosure

Use GitHub Security Advisories or email security details to the project maintainers.

Include:

  • Description of the vulnerability
  • Steps to reproduce
  • Potential impact
  • Suggested fix (if available)
  • Your contact information

Step 2: Acknowledgment

We will acknowledge your report within 48 hours.

Step 3: Investigation

Our team will:

  • Verify the vulnerability
  • Assess severity and impact
  • Develop a fix
  • Test the fix thoroughly

Typical timeline: 5-10 business days

Step 4: Disclosure

Once fixed:

  • Security advisory published
  • Fix released in new version
  • Reporter credited (unless anonymity requested)

Severity Levels

We use the following severity classification:

Critical:

  • Remote code execution
  • Privilege escalation
  • Data exfiltration

High:

  • Authentication bypass
  • SQL injection
  • Local privilege escalation

Medium:

  • Information disclosure
  • Denial of service
  • Path traversal

Low:

  • Minor information leaks
  • Configuration issues

Security Architecture

Local-First = Privacy-First

SuperLocalMemory V2 is designed with security and privacy as foundational principles.

Core Security Principles

1. Zero External Communication

Guarantee: The system makes zero external API calls or network requests.

Implementation:

  • No telemetry
  • No auto-updates
  • No analytics
  • No crash reporting to external services

2. Local-Only Data Storage

All data stored locally:

  • Database: ~/.claude-memory/memory.db
  • Backups: ~/.claude-memory/backups/
  • Profiles: ~/.claude-memory/profiles/
  • Vectors: ~/.claude-memory/vectors/

No cloud synchronization by default.

3. Standard Filesystem Permissions

Security relies on operating system:

  • Unix/Linux: Standard file permissions (chmod 600)
  • macOS: FileVault encryption (user responsibility)
  • Windows: NTFS permissions

Recommendation:

# Secure database file
chmod 600 ~/.claude-memory/memory.db

# Secure entire directory
chmod 700 ~/.claude-memory/

4. No Built-in Authentication

Design decision: Authentication is handled by the OS.

Rationale:

  • Single-user system (runs on local machine)
  • OS-level user authentication sufficient
  • No network exposure = no remote authentication needed

Multi-user environments:

  • Use separate OS user accounts
  • Each user gets isolated ~/.claude-memory/

Data Privacy

What Data is Stored

User-provided data:

  • Memory content (text you add)
  • Tags and metadata
  • Configuration settings

System-generated data:

  • Timestamps
  • Memory IDs
  • Graph cluster information
  • Learned patterns

NOT stored:

  • Personally identifiable information (unless you add it)
  • System information
  • Usage analytics
  • Telemetry

Data Lifecycle

User Input → Local Storage → Processing → Local Output
     ↓
No external transmission at any stage

Data Retention

User control:

  • Keep data forever (default)
  • Delete selectively (per-memory deletion)
  • Reset database (soft/hard reset options)
  • Archive to cold storage (compression system)

Automatic deletion: None. All deletion is user-initiated.

GDPR Compliance

For EU users:

  • Right to access: Full access to all data (it is on your machine)
  • Right to deletion: Use memory-reset commands
  • Right to portability: Export database file
  • Right to rectification: Edit memories directly

No data controller: You control your data completely.


Threat Model

What We Protect Against

1. Data Integrity

Protection:

  • SQLite ACID transactions
  • Automatic backups before destructive operations
  • Schema validation

Mitigations:

  • Crash during write → Database rollback
  • Corrupted data → Restore from backup
  • Accidental deletion → Soft delete with recovery

2. Unauthorized Local Access

Limited protection:

  • Filesystem permissions (user responsibility)
  • No application-level encryption (by design)

Recommendation: Use full-disk encryption (FileVault, LUKS, BitLocker).

3. Code Injection

Protection:

  • Parameterized SQL queries (no string concatenation)
  • Input validation
  • No eval or exec on user input

What We DON'T Protect Against

1. Physical Access

Not protected:

  • Attacker with physical access to machine
  • Attacker with root/admin privileges
  • Forensic data recovery from disk

Mitigation: Use full-disk encryption.

2. Malware on Same Machine

Not protected:

  • Keyloggers
  • Screen capture malware
  • Filesystem-level malware

Mitigation: Standard OS security practices (antivirus, firewall).

3. Supply Chain Attacks

Limited protection:

  • Python standard library only (reduces dependency risk)
  • Open source (code can be audited)
  • No auto-updates (user initiates updates)

Risk: Compromised Python interpreter or OS.


Best Practices

For Users

1. Enable Full-Disk Encryption

macOS:

System Preferences → Security & Privacy → FileVault

Linux:

Use LUKS encryption during installation

Windows:

Control Panel → System and Security → BitLocker

2. Secure Filesystem Permissions

# Lock down .claude-memory directory
chmod 700 ~/.claude-memory/

# Lock down database
chmod 600 ~/.claude-memory/memory.db

# Lock down backups
chmod 700 ~/.claude-memory/backups/

3. Regular Backups

# Manual backup with timestamp
cp ~/.claude-memory/memory.db ~/Backups/memory-backup.db

4. Avoid Storing Secrets

DO NOT store:

  • API keys
  • Passwords
  • Private keys
  • Tokens
  • Personal identification numbers

If you must:

  • Use a dedicated password manager instead
  • Or encrypt sensitive memories manually

For Developers

1. Code Review

Every PR must:

  • Be reviewed by at least one maintainer
  • Pass automated security checks
  • Not introduce external dependencies without justification

2. Input Validation

Validate all user input before processing.

3. Parameterized Queries

Always use parameterized SQL queries to prevent injection attacks.

4. No Secrets in Code

  • Never hardcode API keys
  • Never commit credentials
  • Use .gitignore for sensitive files

Security Updates

Update Notification

How you will be notified:

  1. GitHub Security Advisories
  2. Release notes (CHANGELOG.md)
  3. Project discussions

What to do:

  1. Read the security advisory
  2. Assess impact on your system
  3. Update to latest version
  4. Verify fix is applied

Update Process

# Backup current version
cp -r ~/.claude-memory ~/.claude-memory.backup

# Pull latest code
cd ~/path/to/SuperLocalMemoryV2-repo
git pull origin main

# Reinstall
./install.sh

# Verify version
memory-status

Security Audit

Self-Audit Checklist

For users:

  • Full-disk encryption enabled
  • .claude-memory/ directory has restrictive permissions
  • Regular backups configured
  • No secrets stored in memories
  • Using latest version

For developers:

  • Code reviewed before merge
  • No external dependencies without justification
  • Input validation on user data
  • Parameterized SQL queries
  • No secrets in code or config

Contact

For security issues, please use GitHub Security Advisories or contact project maintainers directly.

Response Time:

  • Acknowledgment: less than 48 hours
  • Initial assessment: less than 5 business days
  • Patch release: Varies by severity (critical: less than 7 days)

Legal

Responsible Disclosure

We appreciate responsible disclosure of security vulnerabilities.

Safe harbor:

  • Good faith security research is welcomed
  • We will not pursue legal action for responsible disclosure
  • We request you do not publicly disclose until patch is available

Disclaimer

SuperLocalMemory V2 is provided "as is" without warranty. See LICENSE for full terms.

User responsibility:

  • Securing your system
  • Enabling encryption
  • Following best practices
  • Keeping software updated

Security is a shared responsibility.

We build secure software. You secure your environment. Together we protect your data.

There aren’t any published security advisories