Skip to content

SECURITY.md

Latest commit

 

History

History
59 lines (41 loc) · 2.15 KB

SECURITY.md

File metadata and controls

59 lines (41 loc) · 2.15 KB

Security Policy

Supported Versions

Version Supported
1.0.x Yes
< 1.0 No

Threat Model

Mnemosyne is a local-first tool that operates on a user's own filesystem. It does not make network calls, host services, or process untrusted input in its default configuration.

Trust boundaries

  • Filesystem access: Mnemosyne reads and writes files within the project root directory and the .mnemosyne/ subdirectory. Path containment is enforced via os.path.realpath() -- symlinks are resolved before validation. Paths outside the project root are rejected.

  • Daemon socket: When running in daemon mode, communication occurs over a Unix domain socket at .mnemosyne/mnemosyne.sock with permissions set to 0600 (owner-only). The daemon validates all ingest paths against the project root before processing.

  • Serialization: All persistent data uses JSON or SQLite. No pickle, eval, marshal, or other unsafe deserialization is used anywhere in the codebase.

  • Database: SQLite with WAL mode. No network-accessible database. No multi-user access model. FTS5 queries are parameterized (no SQL injection).

Out of scope

  • Network security (Mnemosyne makes no network calls in normal operation)
  • Authentication/authorization (single-user local tool)
  • Denial of service (local tool, user controls their own resources)

Known Limitations

  • TOCTOU on path validation: A race condition exists between path resolution and file read. This is inherent to POSIX filesystems and not exploitable in Mnemosyne's single-user context.

  • FTS5 query injection: While FTS5 special characters are escaped, the FTS5 query language is limited and cannot cause data modification. The worst case is unexpected empty results.

Reporting a Vulnerability

Please report security issues to security@castnettechnology.com.

We will acknowledge receipt within 48 hours and provide an initial assessment within 7 days. Critical issues affecting data integrity or path traversal will be patched within 72 hours of confirmation.

Do not open public GitHub issues for security vulnerabilities.