Agent or proxyT1190

Detect known exploits with Suricata rules

When the security community already has a signature for an attack, run it on the host. Thalamus loads Suricata-compatible rules unchanged, across thirty-plus protocols. Tell it what your network is, read the alerts as EVE JSON, and decide what a hit does: alert only, ban the source with ids.enforce_block, or act on it in a smart firewall rule.

Overview

Most attacks are not new. Exploits for published flaws, known C2 frameworks and commodity malware have signatures, written and maintained by the security community, and you may already license them for Suricata.

Thalamus runs those rules inside the agent with no translation step. It parses protocols rather than raw bytes, and writes Suricata-compatible EVE JSON, so the pipelines and dashboards you already have keep working.

When you see this

  1. An exploit for a published flaw is being sprayed across the internet.
  2. You license ET Pro or write your own Suricata rules and want them on every host.
  3. You need detection beyond HTTP: SMB, RDP, DNS, Modbus.

Steps

  1. 01

    Turn on the IDS and load rules

    Thalamus takes Suricata-compatible rules as they are: ET Open, ET Pro or your own. Start with enforce_block off; the shipped configuration recommends that until the rules are tuned.

    /etc/synapse/config.yaml
    ids:
      enabled: true
      capture_mode: xdp
      rule_paths:
        - "/etc/synapse/rules/*.rules"
      enforce_block: false
  2. 02

    Tell it what your network is

    Suricata rules are written against address variables such as HOME_NET. Define them under ids.address_vars. On Kubernetes, the operator's netvars resolver fills them from the cluster instead.

    /etc/synapse/config.yaml
    ids:
      address_vars: { HOME_NET: "10.0.0.0/8" }
    Or, on Kubernetes: synapse-stack values
    operator:
      netvarsResolver: true
  3. 03

    Read the alerts

    Alerts are Suricata-compatible EVE JSON. Thalamus never samples, but the capture layer can drop packets before the engine if traffic outruns the host. That is counted as packets_dropped: if it is not zero, you are not seeing everything.

  4. 04

    Decide what a hit does

    With enforce_block on, a hit bans the offending source through the firewall. The per-flow kernel drop is a separate key, enforce_edge_drop.

    /etc/synapse/config.yaml
    ids:
      enforce_block: true
    DecisionAre the rules tuned for your traffic?

    Yes Turn on enforce_block.

    No Leave it off and act through the smart firewall, where a rule can require a second signal.

  5. 05

    Act on a hit with a rule

    IDS context is a set of rule fields. ids.sids lists the signatures that fired; ids.score runs from 0 to 100, higher is worse. ids.severity runs the other way: 1 is the most severe.

    Smart firewall rule, action drop
    any(ids.sids[*] in {2054155 2047703})
    Smart firewall rule, action drop
    ids.score ge 70 and ml.traffic.c2 ge 50
  6. 06

    Update rules without a restart

    On the XDP IDS worker, rule paths and tuning reload with SIGHUP: the signature engine is rebuilt and swapped, and in-flight flow state is reset. On the AF_PACKET worker, IDS changes need a restart.

    Reload
    kill -HUP $(cat /var/run/synapse.pid)

Common mistakes

  1. Turning on enforce_block before the rules are tuned.
  2. Reading ids.severity > 3 as serious: it selects the least serious alerts.
  3. Ignoring packets_dropped: a host that drops at capture is not inspecting everything.

Escalate when

  • A hit on a C2 signature from inside your network.
  • Repeated hits for an exploit against a service you run.

Where this stops working

  • A signature catches what has been seen before. New tools need behaviour: see the model verdict playbook.
  • Packets dropped by the capture layer are never inspected; packets_dropped says how many.
Run it here

Send the request. Watch where it stops.

Edit the request or any rule and the pipeline runs again. The rules are real wirefilter expressions, evaluated the way the agent evaluates them; the traffic and model scores are samples. It all runs in your browser.

Try this Turn on ids.enforce_block and the IDS bans the source itself. Or leave it off and switch the smart firewall rule from log to drop: the same hit, enforced by a rule you can order and scope.

The request

ingress · agent mode

Every value is a rule field. Change one and the pipeline re-runs.

Connection
Source, resolved by the platform
Fingerprints, from Dendrite
Model scores, sample output from Cortex
Flow behaviour, per source
Payload, what Thalamus inspects
Result Delivered to the application
ResultDelivered to the applicationAllowed, with findings · Logged by Amygdala
  1. 01Capture and fingerprintDendrite · On the wireObserved

    Computed JA4, JA4T for this connection. Capture only: Dendrite never blocks.

  2. 02Access rulesHillock · XDP hook, before routingPassed

    198.51.100.88 (NL, AS64503) is on no access list.

    From the threat feed: known-bad sourcesNo match
    Country, resolved to IP ranges on the platform sideNo match
  3. 03Firewall rulesHillock · TC hook, in the kernelNot used

    No firewall rules in this policy.

  4. 04Intrusion detectionThalamus · On the hostFlagged

    Alert on sid 9000201. IDS context becomes rule fields: ids.alert_count, ids.sids.

    sid 9000201Sample: JNDI lookup in a requestMatched

    Thalamus runs Suricata-compatible signatures across thirty-plus protocols. Here a signature is reduced to one content match.

  5. 05Behavioural modelsCortex · On the host CPUObserved

    Sample scores, 0 to 100: C2 2, scan 3, malicious flow 4. Nothing leaves the machine to be scored.

  6. 06Service graphHippocampus · On the hostNot used

    No workload identity on this flow.

  7. 07Smart firewall rulesAmygdala · Decided on the host, enforced in nftablesFlagged

    Matched a log rule: recorded what it would have done, enforced nothing.

    Matched
  8. 08WAF rulesProxy · In the proxy, after decryptionNot used

    Agent mode does no HTTP proxying, so there is no request for the WAF to read.

  9. 09Content scanningProxy · ClamAV, before forwardingNot used

    No WAF rule selected this request for scanning, so nothing was scanned.

Field names and operators follow the docs: WAF and Smart Firewall. A rule that names an unknown field does not load; a field the request does not carry never matches.

What lands in the event log
{
  "ts": "2026-11-17T09:41:07.212Z",
  "layer": "ids",
  "action": "notice",
  "source": "ids",
  "rule_id": "sid:9000201",
  "rule_msg": "Sample: JNDI lookup in a request",
  "src_ip": "198.51.100.88",
  "src_port": 54321,
  "dst_ip": "198.51.100.10",
  "dst_port": 443,
  "src_country": "NL",
  "fingerprints": {
    "ja4t": "64240_2-4-8-1-3_1460_10",
    "ja4": "t13d1516h2_8daaf6152771_02713d6af862"
  }
}
{
  "ts": "2026-11-17T09:41:07.212Z",
  "layer": "smart_firewall",
  "action": "log",
  "source": "ids",
  "signals": [
    "ids"
  ],
  "rule_id": "s24",
  "src_ip": "198.51.100.88",
  "src_port": 54321,
  "dst_ip": "198.51.100.10",
  "dst_port": 443,
  "src_country": "NL",
  "fingerprints": {
    "ja4t": "64240_2-4-8-1-3_1460_10",
    "ja4": "t13d1516h2_8daaf6152771_02713d6af862"
  }
}

Shaped like the documented event schema; the timestamp, ports, rule ids and ban duration are illustrative.

Now on your own traffic.

A demo runs the same pipeline against real packets, with the engineers who built it.