feat(prober,correlation): attacker fingerprint rotation detection (DEBT-032)

When the prober observes a NEW hash for an
(attacker_uuid, port, probe_type) triple it has seen before — VPS
rotation, SSH server rebuild, TLS cert swap — emit a derived
attacker.fingerprint_rotated event carrying both old and new hash.
Detection is a small library (decnet.correlation.fingerprint_rotation)
called inline from the prober at each of the three emit sites
(JARM/HASSH/TCPFP). No new daemon. New AttackerFingerprintState table
holds per-triple last-hash state; Attacker.rotation_count and
Attacker.last_rotation_at are stamped on every diff. Library is sync,
fully unit-tested via injected publish_fn / syslog_fn callbacks.
This commit is contained in:
2026-05-03 05:12:51 -04:00
parent dcd558fd91
commit 6c6f97e840
8 changed files with 687 additions and 18 deletions

View File

@@ -0,0 +1,236 @@
"""Unit tests for ``decnet.correlation.fingerprint_rotation``.
Pure library: in-memory SQLite + sync Session + collected callback
calls. No prober, no bus, no async. Each test seeds an Attacker row,
calls ``record_fingerprint``, asserts on the returned outcome + the
side-effects (state row, Attacker stamp, callback invocations).
"""
from __future__ import annotations
from datetime import datetime, timedelta, timezone
import pytest
from sqlalchemy.engine import Engine
from sqlmodel import Session, SQLModel, create_engine, select
from decnet.correlation.fingerprint_rotation import (
record_fingerprint,
RotationOutcome,
)
from decnet.web.db.models import (
Attacker,
AttackerFingerprintState,
)
@pytest.fixture
def engine() -> Engine:
eng = create_engine("sqlite://", connect_args={"check_same_thread": False})
SQLModel.metadata.create_all(eng)
return eng
@pytest.fixture
def now() -> datetime:
return datetime(2026, 5, 3, 12, 0, 0, tzinfo=timezone.utc)
def _seed_attacker(session: Session, ip: str = "1.2.3.4") -> Attacker:
a = Attacker(
uuid="attacker-uuid-1",
ip=ip,
first_seen=datetime.now(timezone.utc),
last_seen=datetime.now(timezone.utc),
)
session.add(a)
session.commit()
session.refresh(a)
return a
class _Recorder:
"""Capture (event_type, payload) tuples from publish_fn / syslog_fn."""
def __init__(self) -> None:
self.calls: list[tuple[str, dict]] = []
def __call__(self, event_type: str, payload: dict) -> None:
self.calls.append((event_type, payload))
def test_no_attacker_row_returns_noop(engine, now):
publish, syslog = _Recorder(), _Recorder()
with Session(engine) as session:
outcome = record_fingerprint(
session,
attacker_ip="9.9.9.9",
port=22,
probe_type="hassh",
new_hash="abc",
ts=now,
publish_fn=publish,
syslog_fn=syslog,
)
assert outcome.kind == "no_attacker_row"
assert publish.calls == []
assert syslog.calls == []
with Session(engine) as session:
rows = session.exec(select(AttackerFingerprintState)).all()
assert rows == []
def test_first_sighting_creates_state_row_no_event(engine, now):
publish, syslog = _Recorder(), _Recorder()
with Session(engine) as session:
_seed_attacker(session)
outcome = record_fingerprint(
session,
attacker_ip="1.2.3.4",
port=22,
probe_type="hassh",
new_hash="hash-1",
ts=now,
publish_fn=publish,
syslog_fn=syslog,
)
assert outcome.kind == "first_sighting"
assert outcome.old_hash is None
assert outcome.new_hash == "hash-1"
assert outcome.rotation_count == 0
assert publish.calls == []
assert syslog.calls == []
with Session(engine) as session:
rows = session.exec(select(AttackerFingerprintState)).all()
assert len(rows) == 1
assert rows[0].last_hash == "hash-1"
assert rows[0].rotation_count == 0
a = session.exec(select(Attacker)).one()
assert a.rotation_count == 0
assert a.last_rotation_at is None
def test_unchanged_hash_bumps_last_seen_no_event(engine, now):
publish, syslog = _Recorder(), _Recorder()
later = now + timedelta(minutes=10)
with Session(engine) as session:
_seed_attacker(session)
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="hash-1", ts=now,
)
outcome = record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="hash-1", ts=later,
publish_fn=publish, syslog_fn=syslog,
)
assert outcome.kind == "unchanged"
assert publish.calls == []
assert syslog.calls == []
with Session(engine) as session:
row = session.exec(select(AttackerFingerprintState)).one()
# SQLite strips tzinfo on round-trip; compare naive values.
assert row.last_seen.replace(tzinfo=timezone.utc) == later
assert row.rotation_count == 0
def test_rotated_emits_event_and_stamps_attacker(engine, now):
publish, syslog = _Recorder(), _Recorder()
later = now + timedelta(hours=1)
with Session(engine) as session:
_seed_attacker(session)
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="hash-1", ts=now,
)
outcome = record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="hash-2", ts=later,
publish_fn=publish, syslog_fn=syslog,
)
assert outcome.kind == "rotated"
assert outcome.old_hash == "hash-1"
assert outcome.new_hash == "hash-2"
assert outcome.rotation_count == 1
assert len(publish.calls) == 1
assert len(syslog.calls) == 1
event_type, payload = publish.calls[0]
assert event_type == "attacker.fingerprint_rotated"
assert payload["attacker_uuid"] == "attacker-uuid-1"
assert payload["attacker_ip"] == "1.2.3.4"
assert payload["port"] == 22
assert payload["probe_type"] == "hassh"
assert payload["old_hash"] == "hash-1"
assert payload["new_hash"] == "hash-2"
assert payload["rotation_count"] == 1
assert payload["ts"] == later.isoformat()
with Session(engine) as session:
a = session.exec(select(Attacker)).one()
assert a.rotation_count == 1
assert a.last_rotation_at is not None
assert a.last_rotation_at.replace(tzinfo=timezone.utc) == later
row = session.exec(select(AttackerFingerprintState)).one()
assert row.last_hash == "hash-2"
assert row.rotation_count == 1
def test_three_probe_types_independent(engine, now):
with Session(engine) as session:
_seed_attacker(session)
for ptype in ("jarm", "hassh", "tcpfp"):
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type=ptype,
new_hash=f"{ptype}-1", ts=now,
)
with Session(engine) as session:
rows = session.exec(select(AttackerFingerprintState)).all()
assert {r.probe_type for r in rows} == {"jarm", "hassh", "tcpfp"}
assert {r.last_hash for r in rows} == {"jarm-1", "hassh-1", "tcpfp-1"}
def test_two_ports_same_probe_type_independent(engine, now):
with Session(engine) as session:
_seed_attacker(session)
for port in (22, 2222):
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=port, probe_type="hassh",
new_hash=f"hash-{port}", ts=now,
)
with Session(engine) as session:
rows = session.exec(select(AttackerFingerprintState)).all()
assert {r.port for r in rows} == {22, 2222}
def test_multiple_rotations_increment_counter(engine, now):
publish = _Recorder()
with Session(engine) as session:
_seed_attacker(session)
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="h1", ts=now, publish_fn=publish,
)
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="h2", ts=now + timedelta(minutes=5), publish_fn=publish,
)
record_fingerprint(
session,
attacker_ip="1.2.3.4", port=22, probe_type="hassh",
new_hash="h3", ts=now + timedelta(minutes=10), publish_fn=publish,
)
assert len(publish.calls) == 2 # first call was first_sighting (no event)
with Session(engine) as session:
a = session.exec(select(Attacker)).one()
assert a.rotation_count == 2
row = session.exec(select(AttackerFingerprintState)).one()
assert row.rotation_count == 2
assert row.last_hash == "h3"

View File

@@ -0,0 +1,142 @@
"""Integration test: prober phase functions invoke the rotation recorder.
The prober worker constructs the recorder closure at startup; here we
verify that ``_probe_cycle`` threads a recorder through to JARM / HASSH
/ TCPFP phases and that the recorder gets the (ip, port, probe_type,
hash) tuple it expects. The library itself is unit-tested separately.
"""
from __future__ import annotations
from pathlib import Path
from unittest.mock import MagicMock, patch
from decnet.prober.worker import _probe_cycle
@patch("decnet.prober.worker.fetch_leaf_cert", return_value=None)
@patch("decnet.prober.worker.tcp_fingerprint", return_value=None)
@patch("decnet.prober.worker.hassh_server", return_value=None)
@patch("decnet.prober.worker.jarm_hash")
def test_jarm_phase_calls_recorder(
mock_jarm: MagicMock,
_mock_hassh: MagicMock,
_mock_tcpfp: MagicMock,
_mock_cert: MagicMock,
tmp_path: Path,
):
mock_jarm.return_value = "c0c" * 10 + "a" * 32
log_path = tmp_path / "decnet.log"
json_path = tmp_path / "decnet.json"
rec_calls: list[tuple] = []
recorder = lambda ip, port, ptype, h: rec_calls.append((ip, port, ptype, h)) # noqa: E731
_probe_cycle(
{"10.0.0.5"}, {},
[443], [], [],
log_path, json_path,
timeout=1.0,
publish_fn=None,
record_rotation=recorder,
)
assert rec_calls == [("10.0.0.5", 443, "jarm", "c0c" * 10 + "a" * 32)]
@patch("decnet.prober.worker.fetch_leaf_cert", return_value=None)
@patch("decnet.prober.worker.tcp_fingerprint", return_value=None)
@patch("decnet.prober.worker.hassh_server")
@patch("decnet.prober.worker.jarm_hash", return_value="")
def test_hassh_phase_calls_recorder(
_mock_jarm: MagicMock,
mock_hassh: MagicMock,
_mock_tcpfp: MagicMock,
_mock_cert: MagicMock,
tmp_path: Path,
):
mock_hassh.return_value = {
"hassh_server": "deadbeef",
"banner": "SSH-2.0-OpenSSH_9.2",
"kex_algorithms": "x",
"encryption_s2c": "x",
"mac_s2c": "x",
"compression_s2c": "x",
}
log_path = tmp_path / "decnet.log"
json_path = tmp_path / "decnet.json"
rec_calls: list[tuple] = []
recorder = lambda ip, port, ptype, h: rec_calls.append((ip, port, ptype, h)) # noqa: E731
_probe_cycle(
{"10.0.0.5"}, {},
[], [22], [],
log_path, json_path,
timeout=1.0,
publish_fn=None,
record_rotation=recorder,
)
assert rec_calls == [("10.0.0.5", 22, "hassh", "deadbeef")]
@patch("decnet.prober.worker.fetch_leaf_cert", return_value=None)
@patch("decnet.prober.worker.tcp_fingerprint")
@patch("decnet.prober.worker.hassh_server", return_value=None)
@patch("decnet.prober.worker.jarm_hash", return_value="")
def test_tcpfp_phase_calls_recorder(
_mock_jarm, _mock_hassh, mock_tcpfp, _mock_cert, tmp_path: Path,
):
mock_tcpfp.return_value = {
"tcpfp_hash": "tcpfp-hash-1",
"tcpfp_raw": "raw",
"ttl": 64,
"window_size": 65535,
"df_bit": True,
"mss": 1460,
"window_scale": 7,
"sack_ok": True,
"timestamp": True,
"options_order": "MSS,SACK,TS,NOP,WS",
"tos": 0,
"dscp": 0,
"ecn": 0,
"server_isn": 0,
}
log_path = tmp_path / "decnet.log"
json_path = tmp_path / "decnet.json"
rec_calls: list[tuple] = []
recorder = lambda ip, port, ptype, h: rec_calls.append((ip, port, ptype, h)) # noqa: E731
_probe_cycle(
{"10.0.0.5"}, {},
[], [], [22],
log_path, json_path,
timeout=1.0,
publish_fn=None,
record_rotation=recorder,
)
assert rec_calls == [("10.0.0.5", 22, "tcpfp", "tcpfp-hash-1")]
@patch("decnet.prober.worker.fetch_leaf_cert", return_value=None)
@patch("decnet.prober.worker.tcp_fingerprint", return_value=None)
@patch("decnet.prober.worker.hassh_server", return_value=None)
@patch("decnet.prober.worker.jarm_hash")
def test_recorder_optional_no_crash_when_none(
mock_jarm: MagicMock,
_mock_hassh: MagicMock,
_mock_tcpfp: MagicMock,
_mock_cert: MagicMock,
tmp_path: Path,
):
"""record_rotation=None must keep the prober's pre-DEBT-032 behavior."""
mock_jarm.return_value = "c0c" * 10 + "a" * 32
_probe_cycle(
{"10.0.0.5"}, {},
[443], [], [],
tmp_path / "decnet.log", tmp_path / "decnet.json",
timeout=1.0,
publish_fn=None,
record_rotation=None,
)
# No error, probe completes.