Triggers
Start a flow from a new alert, a case event, a triage verdict, a schedule, or a person, and narrow alert and case triggers with an event match.
Choose a trigger to decide when a flow runs. Only live flows fire on their trigger. Any flow that isn't archived can also run manually.
Trigger types
In the flow editor, pick a trigger under Triggers in the block palette, or select the trigger and change Trigger type in the inspector.
| Trigger | on value | Fires when | Event data in the run |
|---|---|---|---|
| Alert committed | alert | A new alert is created | flow_input.alert |
| Case event | case | A case is created, closed, or reopened, its status changes, or it is set to or cleared from pending | flow_input.case |
| Case triage verdict | case.triage_verdict | AI triage records a verdict on a case | flow_input.case, verdict, confidence, alert_ids, top_entities |
| Schedule | schedule | A cron schedule comes due | None |
| Manual | manual | Someone runs it from the flow editor or the API | Only the inputs they provide |
Through the API you can also use case_created, which fires only when a case is created.
One run per event
An event starts at most one run of each matching flow. If ten flows match an alert, nano starts ten runs, one for each flow. The same alert never starts a second run of the same flow. Each case event and each triage verdict is its own event.
During an alert storm, nano starts one run for every matching alert for every matching flow. Runs queue and drain at the executor's concurrency limit. See How runs behave under load.
Narrow a trigger with an event match
Alert, case, and triage verdict triggers take an optional Event match. An empty match fires on every event of that type. The match is checked against the trigger's event data before nano starts a run.
Match grammar
A match is one or more clauses joined by &&. Each clause is one of these:
| Clause | Example | Meaning |
|---|---|---|
flow_input.<path> = "value" (or ==) | flow_input.alert.rule_name = "Suspicious PowerShell" | The whole value equals the literal, ignoring case and surrounding spaces |
flow_input.<path> in ["a", "b"] | flow_input.case.status in ["open", "in_progress"] | The value equals one of the literals |
flow_input.alert.severity <op> "<severity>" | flow_input.alert.severity >= "high" | Severity comparison with >, >=, <, or <= |
flow_input.case.severity <op> "<severity>" | flow_input.case.severity > "medium" | The same, on the case |
flow_input.confidence <op> <number> | flow_input.confidence >= 0.9 | Triage confidence between 0 and 1, with >, >=, <, <=, or = |
Severities are ordered informational < low < medium < high < critical.
Rules:
- At most 8 clauses, and at most 8,192 bytes.
||(or) isn't supported. For alternatives on one field, usein [...]. For "high or above", use>= "high".- A clause can read only
flow_inputfields. - Fail closed: a missing or unknown severity never matches an ordering, a missing confidence never matches a threshold, and a match nano can't parse never matches.
Examples:
flow_input.alert.severity >= "high" && flow_input.alert.entities.host in ["dc01", "dc02"]flow_input.verdict = "malicious" && flow_input.confidence >= 0.8flow_input.case.event = "closed"Filter by severity and rule
Through the API, an alert trigger also takes severity (a list of severities) and rule_ids (at most 64 detection rule IDs). A case_created trigger takes severity. These filters are checked as the alert or case is written: when no live flow wants an event, it never enters the trigger queue. An event match is checked later, when nano reads the queue, so an event that fails only its match still takes a place in the queue.
{"on": "alert", "severity": ["high", "critical"], "rule_ids": ["<rule id>"], "match": "flow_input.alert.entities.user in [\"svc_backup\"]"}Trigger payloads
Steps and event matches read the trigger's data under flow_input.
Alert
| Field | Contents |
|---|---|
flow_input.alert.id | Alert ID |
flow_input.alert.rule_id, rule_name | The detection rule |
flow_input.alert.title | The rule name. Without a rule, the alert kind followed by "alert", for example "Detection alert" |
flow_input.alert.tags | The detection rule's tags |
flow_input.alert.severity | Alert severity |
flow_input.alert.kind | Alert kind |
flow_input.alert.created_at | When the alert was created |
flow_input.alert.entities.<name> | The first value found in the alert's matched events for host, user, src_ip, dest_ip, process, command_line, hash, or domain, or null |
nano resolves entities from OCSF fields by event class, for example actor.user.name on endpoint events. It reads the first 50 matched events and skips values longer than 1,024 bytes.
Case event
flow_input.case holds id, case_number, severity, status, priority, and event. event is created, closed, reopened, status_changed, pending_set, or pending_cleared. For a change, previous_status holds the status before it. Case payloads carry no title or description.
Case triage verdict
| Field | Contents |
|---|---|
flow_input.case | id, case_number, status, severity |
flow_input.verdict | malicious, suspicious, benign, or inconclusive |
flow_input.confidence | 0 to 1, or null |
flow_input.alert_ids | Up to 100 alert IDs on the case |
flow_input.top_entities | Up to 10 case entities, each with type and value |
Schedules
A schedule uses a 7-field cron in UTC: sec min hour day month weekday year. In the inspector, set Every (minutes) or type the cron directly. The inspector shows the next run time.
0 0 */6 * * * *- Runs must be at least 5 minutes apart. nano checks the next 8 fire times and refuses a cron that runs more often.
- A schedule doesn't catch up. If nano was down or SOAR was off when a run was due, the next run is the next occurrence from now.
- A scheduled run carries no event data.
- If a live schedule flow can no longer start a run, for example because it no longer compiles or its owner lost a permission, nano returns it to
draft.
Who a triggered run acts for
A triggered run acts for the person who last saved the flow, and nano checks that person's permissions before every step. It also checks that this person can see the alert or case that fired the trigger. If they can't, nano skips the flow for that event.
What's next
Flows
Build a flow from actions, searches, notifications, Judge steps, approval gates, conditions, and loops; test it safely; then activate it and follow its runs.
Actions and connections
Install signed action packs, connect them to your security tools with write-only credentials, run a single action, and control where SOAR may send requests.