Skip to content

Security: Omkar-git-hub/msk-project-intelligence

Security

docs/security.md

MSK Security Model

1. Principles

  1. Zero Secret Persistence: Under no circumstance are secret keys, passwords, bearer tokens, private keys, or credentials stored inside .msk/msk.db or .msk/*.json.
  2. Local Analysis Only: Core analysis operates offline without external network communication.
  3. Boundary Integrity: In future AI phases, context passed to LLMs must pass through strict policy filtering and redaction.
  4. Symlink Traversal Protection: The file scanner rejects and skips symlinks that point outside the project boundary or create directory recursion loops.
  5. Safe Subprocess Execution: Any interaction with native Git commands uses parameterized lists rather than raw shell strings to prevent injection.

There aren't any published security advisories