Skip to content

SECURITY.md

Latest commit

 

History

History
85 lines (61 loc) · 3.08 KB

SECURITY.md

File metadata and controls

85 lines (61 loc) · 3.08 KB
Raw

Security Policy

SuperLocalMemory V3 Security

Supported Versions

Version Supported
3.7.x Yes
3.6.x Security fixes only
< 3.6 No

Reporting Vulnerabilities

Do NOT open public issues for security vulnerabilities.

Email: admin@superlocalmemory.com

Include:

  • Description of the vulnerability
  • Steps to reproduce
  • Impact assessment
  • Suggested fix (if any)

We will respond within 48 hours and provide a fix timeline within 7 days.

Security Architecture

Mode A (Zero-LLM, Local-Only)

  • All data stored at ~/.superlocalmemory/
  • Zero cloud API calls during store/recall operations
  • No telemetry, analytics, or phone-home code
  • SQLite with WAL mode for data integrity

Authentication

  • Optional API key authentication for dashboard/API access
  • HMAC-SHA256 timing-safe comparison
  • Rate limiting: 30 writes/min, 120 reads/min

Data Protection

  • Parameterized SQL queries throughout (no SQL injection)
  • XSS protection via escapeHtml() in all UI rendering
  • CSRF protection via token-based auth (no cookies)
  • Security headers: X-Frame-Options, CSP, X-Content-Type-Options
  • CORS whitelist with credential control

Model Supply Chain (Untrusted Checkpoints)

SuperLocalMemory loads local embedding, reranker, and compression models via Hugging Face. Two risks apply:

  • trust_remote_code: since v3.7.9, trust_remote_code=True is passed only for an internal allowlist of pinned models (the nomic-embed family, which requires custom modeling code). Any other model — including one swapped into config through a write path — loads with trust_remote_code=False and cannot execute repository code at load time.
  • CVE-2025-14926 (transformers, SEW tokenizer): arbitrary code execution when loading a crafted SEW tokenizer config. No upstream patch exists as of 2026-07-20. SLM never loads checkpoints automatically; it will be upgraded as soon as a patch ships.

Only install models from sources you trust. A malicious checkpoint can execute code under your user account at load time regardless of these mitigations.

Compliance

  • GDPR Article 15 (right to access): full data export
  • GDPR Article 17 (right to erasure): complete erasure including learning data
  • EU AI Act data sovereignty: Mode A keeps all data local
  • Tamper-proof audit trail with SHA-256 hash chain

Dependencies

Run npm audit and pip audit regularly. Report any findings.

Research Foundation

The security architecture is formally documented across three peer-reviewed papers:

  • Paper 1 (arXiv:2603.02240): Bayesian trust defense, OWASP-aligned memory poisoning protection
  • Paper 2 (arXiv:2603.14588): Information-geometric foundations, cellular sheaf cohomology for contradiction detection
  • Paper 3 (arXiv:2604.04514): Trust-weighted forgetting, compliance audit trails, FRQAD mixed-precision integrity

Part of Qualixar | Author: Varun Pratap Bhardwaj