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
fix(ci): close the six breakers blocking both required Python checks on main (#1215)
Beenu Arora4067cbaf4dbc
diff --git a/apps/docs/docs/compliance/evidence-pack.md b/apps/docs/docs/compliance/evidence-pack.mdindex b8f26262e..c105d9e43 100644--- a/apps/docs/docs/compliance/evidence-pack.md+++ b/apps/docs/docs/compliance/evidence-pack.md@@ -33,7 +33,7 @@ and says so.| Prompt-injection resistance | Three adversarial corpora: 32 payloads scanned in isolation and 71 incident pairs whose payloads are placed in attacker-controllable fields, both against a clean twin with a recall floor and a false-positive ceiling, plus 34 held-out pairs authored after the guard was frozen and carrying no floor at all, on which the guard measures 7.1% | `services/agents/tests/adversarial/` || Claim-to-gate traceability | Every product claim mapped to the CI job that fails when it stops being true | `docs/audit/CLAIM_TO_GATE_MATRIX.md` |-The last row is the one worth reading first. It is the honest index: 289 rows+The last row is the one worth reading first. It is the honest index: 292 rows`GATED`, 1 `PARTIAL`, and a ratchet that refuses to let a row sit at`NO GATE`. Re-read the figure from the table rather than from here — a countcopied into prose goes stale in silence, which is whydiff --git a/apps/docs/docs/quickstart.md b/apps/docs/docs/quickstart.mdindex d78c486ea..e8fb12450 100644--- a/apps/docs/docs/quickstart.md+++ b/apps/docs/docs/quickstart.md@@ -237,7 +237,7 @@ does not.| *(default — CORE)* | **postgres** (5432) · **redis** (6379) · **zookeeper** · **kafka** (9092) · **qdrant** (6333) · **litellm** (4000, LLM gateway) · **ollama** (11434, local model) · **ollama-pull** (one-shot model fetch) · **ingest-worker** (8081, Go) · **fusion** (8003) · **api** (8000) · **agents** (8001) · **threatintel** (8005) · **realtime** (8086) · **web** (3000) · **connectors** (8088) · **actions** (8002) || `full` | **clickhouse** (8123/9000) · **neo4j** (7474/7687) · **opensearch** (9200) · **ueba** (8007) · **enrichment** (8080, Go) · **kafka-ui** (8090) || `extras` | **honeytokens** (8008) · **purple-team** (8006) |-| `chatops` | **slack-bot** (8009) |+| `chatops` | **slack-bot** (8009) · **teams-bot** (8010) || `monitoring` | **prometheus** (9091) · **grafana** (3001) · **alertmanager** (9094) · **tempo** (3200) · **otel-collector** (4317/4318) || `osquery` | **osquery-tls** (8091) |@@ -248,7 +248,7 @@ federated search defaults on, so the capability was enabled and aimed at ahostname that does not resolve; the measured cost of moving them is 124.8 MiB([ADR-0007](https://github.com/beenuar/AiSOC/blob/main/docs/decisions/0007-connectors-and-actions-in-core.md)).They are now in no profile, which means every profile run includes them, so-`chatops` still gets `actions` for `slack-bot`. Named profiles compose, so+`chatops` still gets `actions` for both chat bots. Named profiles compose, so`docker compose --profile full --profile extras up -d` is valid; it just doesnot wait for health the way `make up-full` does.diff --git a/docker-compose.prod.yml b/docker-compose.prod.ymlindex 7bd2dca1d..e540d262e 100644--- a/docker-compose.prod.yml+++ b/docker-compose.prod.yml@@ -449,6 +449,12 @@ services:condition: service_completed_successfullyports: !reset []+ teams-bot:+ depends_on:+ preflight-secrets:+ condition: service_completed_successfully+ ports: !reset []+tempo:depends_on:preflight-secrets:diff --git a/services/teams-bot/app/notify.py b/services/teams-bot/app/notify.pyindex 9837ff868..2335e9938 100644--- a/services/teams-bot/app/notify.py+++ b/services/teams-bot/app/notify.py@@ -79,7 +79,9 @@ def _authorized(supplied: str | None) -> bool:expected = os.environ.get("AISOC_INTERNAL_TOKEN", "").strip()if not expected:return False- return bool(supplied) and hmac.compare_digest(supplied.strip(), expected)+ if not supplied:+ return False+ return hmac.compare_digest(supplied.strip(), expected)@router.post("/approval-card", status_code=status.HTTP_202_ACCEPTED)diff --git a/services/teams-bot/tests/test_notify.py b/services/teams-bot/tests/test_notify.pyindex 9119b534b..9b863092a 100644--- a/services/teams-bot/tests/test_notify.py+++ b/services/teams-bot/tests/test_notify.py@@ -44,6 +44,26 @@ def test_an_unset_internal_token_refuses_everything(monkeypatch: pytest.MonkeyPaassert response.status_code == 401[email protected]("supplied", ["", " ", "wrong", "tok-with-suffix"])+def test_a_token_that_is_not_the_configured_one_is_refused(client: TestClient, supplied: str) -> None:+ """Pins each branch of the comparison, including the empty header.++ `_authorized` reached these cases through `bool(supplied) and ...`,+ which does not narrow `str | None` for a type checker. Rewriting it as+ an early return does, and these are the inputs that would show a+ behaviour change if the rewrite had introduced one -- an empty or+ whitespace header in particular, which the truthiness test rejected+ before reaching the compare and the early return rejects at the same+ point.+ """+ response = client.post(+ "/internal/approval-card",+ json={"action": _ACTION},+ headers={"X-AiSOC-Internal-Token": supplied},+ )+ assert response.status_code == 401, f"{supplied!r} was accepted as the internal token"++def test_no_configured_webhook_is_reported_not_failed(client: TestClient, monkeypatch: pytest.MonkeyPatch) -> None:"""202 with ``posted: false``.