Open-source AI Security Operations Center: alert fusion, LLM-agent triage, MITRE ATT&CK investigation, and a replayable decision ledger for every agent step. Self-hostable, runs with no API keys, MIT licensed. Ships an MCP server for Claude, Cursor and Continue.
Python agentic-aiai-securityalert-triageblue-teamcybersecuritydetection-engineeringincident-responselanggraphllm-agentsmcpmitre-attackmodel-context-protocolpurple-teamsecurity-operationsself-hostedsiemsoarsocthreat-detectionthreat-intelligence
"""verify_chain_tail: truncation of the newest audit rows must be visible."""
from app.services.audit_hash import verify_chain_tail
def _rows(n):
return [{"chain_index": i, "entry_hash": f"h{i}"} for i in range(n)]
def test_intact_tail_is_not_a_break():
assert verify_chain_tail(_rows(3), {"next_index": 3, "head_hash": "h2"}) is None
def test_deleting_the_newest_row_is_a_break():
# Observed on v17.1: rows 0..2 written, row 2 deleted with the immutability
# trigger disabled, /health/audit-chain still reported replay_intact=true.
brk = verify_chain_tail(_rows(2), {"next_index": 3, "head_hash": "h2"})
assert brk is not None and brk["missing"] == 1
def test_deleting_every_row_is_a_break():
brk = verify_chain_tail([], {"next_index": 3, "head_hash": "h2"})
assert brk is not None and brk["missing"] == 3
def test_append_after_the_head_was_read_is_not_a_break():
assert verify_chain_tail(_rows(4), {"next_index": 3, "head_hash": "h2"}) is None
def test_replaced_newest_row_is_a_break():
rows = _rows(3)
rows[-1]["entry_hash"] = "forged"
brk = verify_chain_tail(rows, {"next_index": 3, "head_hash": "h2"})
assert brk is not None and brk["missing"] == 0
def test_no_head_yet_is_not_a_break():
assert verify_chain_tail([], None) is None
assert verify_chain_tail([], {"next_index": 0, "head_hash": None}) is None
def test_legacy_unindexed_rows_are_ignored():
rows = [{"chain_index": None, "entry_hash": "x"}, *_rows(2)]
assert verify_chain_tail(rows, {"next_index": 2, "head_hash": "h1"}) is None