Skip to content

SECURITY.md

Latest commit

 

History

History
29 lines (25 loc) · 1.34 KB

SECURITY.md

File metadata and controls

29 lines (25 loc) · 1.34 KB

Security

Local-first intent

  • Intel-Lint is designed for local analysis of sensitive CTI reports.
  • Default API bind is loopback (127.0.0.1) via python scripts/run.py api or python scripts/run.py app.
  • Recommended engine is local Ollama (ENGINE=ollama, OLLAMA_URL=http://localhost:11434).

Data storage

  • Runtime data is stored in the user data directory (not the repo by default):
    • Windows: %LOCALAPPDATA%\\intel-lint
    • macOS: ~/Library/Application Support/intel-lint
    • Linux: ${XDG_DATA_HOME:-~/.local/share}/intel-lint
  • Subfolders:
    • outputs/latest/ generated outputs (claims.json, annotated.md, rewrite.md)
    • logs/app.log metadata logs (timestamps, engine, output path)
    • config/settings.json local runtime settings
    • outputs/cache/ deterministic cache artifacts
    • debug/ verbose debug artifacts (only when VERBOSE_LOGGING=1)

Wipe local data

  • Delete the user data directory above to remove local outputs, logs, cache, and settings.
  • Optional: remove repo-local .env if used for developer overrides.

Verifying local-only behavior (Ollama mode)

  • Set:
    • ENGINE=ollama
    • OLLAMA_URL=http://localhost:11434
  • Run the app and confirm network activity is only to localhost (127.0.0.1:11434 and local app ports).
  • Keep VERBOSE_LOGGING=0 (default) unless you explicitly need debug artifacts.