secubox-deb/common/secubox_core/classifiers
CyberMind-FR e67baf4cd7 feat(secubox-core+5modules+toolbox): Phase 2c receiving modules enrichment (ref #490)
Phase 2b (#488/#489) wired mitm addons to 5 receiving modules but events
were persisted raw. This phase implements actual enrichment in each module
via the enrich_hook param, plus aggregator merge so reports show meaningful
data (YouTube/Signal/iPhone/Safari/JA4 fingerprints) instead of raw bytes.

  - host_app.py  : 60+ host patterns -> app + category + emoji
  - cookie.py    : 40+ cookie tracker patterns -> provider + category + emoji
  - avatar.py    : UA + Client Hints -> device + browser + OS + emoji
  - ja4.py       : NEW. Deterministic JA4-style fingerprint hash from
                   cipher_suites + alpn + extensions. Lookup table for
                   known JA4 fingerprints (empty for now, Phase 3 will
                   populate). 12-char hex (SHA256 truncated).

The host_app/cookie/avatar are copies of the secubox-toolbox classifiers
moved to secubox-core so all 5 receiving modules can import them. The
secubox-toolbox-local ones stay as-is to avoid breaking changes — Phase 3
will consolidate.

  - secubox-dpi             : host/SNI -> {app, category, emoji}
  - secubox-cookies         : cookie names -> {providers{}, categories{}}
  - secubox-avatar          : UA + CH -> {device, browser, os_label}
  - secubox-threat-analyst  : ClientHello -> {ja4_fingerprint, known_client}
  - secubox-soc             : indicators -> {total_weight, band, kinds}

Each is ~25 lines, called by mount_ingest_routes BEFORE persistence.
Enriched output joins the raw event under the 'enriched' key.

_pull_mitm_module_events() now also calls _summarize_enriched(kind, events)
to consolidate per-module enrichment :

  - dpi             : top_apps[] aggregated from enriched.app counts
  - cookies         : top_providers[] from enriched.providers + tracker_total
  - avatar          : devices{} + browsers{} from enriched.{device,browser}
  - threat-analyst  : top_fingerprints[] grouped by JA4 hash
  - soc             : total_weight + max_band + indicator_kinds

These appear under mitm_modules.<kind>.enriched_summary in the /report JSON.

POST realistic payloads to all 5 sockets :
  - YouTube host -> dpi/enriched.app = 'YouTube' (streaming)
  - GA + FB Pixel cookies -> cookies/providers : GA x3, FB x1, total 4
  - iPhone Safari UA -> avatar/device='iPhone' (📱 iOS 17.4) + Safari (🧭)
  - facebook.com ClientHello -> threat-analyst/ja4 = '7175ee3a68f0'
  - 2 indicators (weight 15+25) -> soc/band='medium', total=40, kinds=[dga, suspicious]

  - secubox-dpi : call live nDPI/netifyd socket (currently pattern-match only)
  - secubox-threat-analyst : implement full FoxIO JA4 string format (currently
    deterministic SHA256 trunc which is JA4-like but not the canonical format)
  - secubox-soc : threat-intel feed lookup (currently just sums static weights)
  - secubox-avatar : screen/timing fingerprinting via WebGL hash (currently UA only)
  - Reports (PDF + HTML) : surface mitm_modules.enriched_summary in the report UI
2026-06-13 08:00:18 +02:00
..
__init__.py feat(secubox-core+5modules+toolbox): Phase 2c receiving modules enrichment (ref #490) 2026-06-13 08:00:18 +02:00
avatar.py feat(secubox-core+5modules+toolbox): Phase 2c receiving modules enrichment (ref #490) 2026-06-13 08:00:18 +02:00
cookie.py feat(secubox-core+5modules+toolbox): Phase 2c receiving modules enrichment (ref #490) 2026-06-13 08:00:18 +02:00
host_app.py feat(secubox-toolbox): DPI classifier 57→147 patterns + fix R3 metrics aggregation (ref #496) 2026-06-06 06:55:10 +02:00
ja4.py feat(secubox-core+5modules+toolbox): Phase 2c receiving modules enrichment (ref #490) 2026-06-13 08:00:18 +02:00
security_quality.py feat(secubox-core+toolbox): Phase 3 transparency layer (whitelist + analysis-status + quality scoring) (ref #492) 2026-06-05 08:13:45 +02:00