When a training organisation says its security “feels busy but not safe”, the complaint usually hides a specific fault in how it handles information about attackers. This page works like a maintenance troubleshooting card: it lists the symptoms you may recognise, the likely cause behind each, and the fix to try first. The subject is ThreatOps, the practice of gathering intelligence on adversaries and hunting for the ones your alarms missed, as laid out in the Ops Atlas write-up on ThreatOps.

Symptom: we subscribe to several feeds, yet nothing seems to change

Likely cause: the feeds arrive with no question attached. The Atlas lists the belief that extra subscriptions bring extra safety among its common errors; volume with no purpose merely adds clutter.

Fix: write down five to eight priority questions, each aimed at a lone question with a single decision attached, and assign every feed a job against them. A feed that supports no question is a candidate for cancellation. Review the list every quarter so it follows your real operations, such as new simulator equipment or a larger student intake, rather than last year’s worries.

Symptom: a firewall started rejecting legitimate users

Likely cause: a feed was wired straight into blocking without measurement. The Atlas describes this exact accident, in which a listed address belonged to shared infrastructure that thousands of normal sites rely on. The FIRST guidance it quotes says feeds suit flagging and alerting, not unattended blocking.

Fix: set entry conditions for any automatic block. In the Atlas’s example, blocking is permitted when the source earns a top-tier grade, the claim scores as credible, the feed has been right at least nine times in ten across a quarter, and the entry sits on none of your allowlists. Let anyone on call reverse a block, have the intelligence owner review it within an hour, and count reversals monthly.

Symptom: old warnings keep firing for months

Likely cause: indicators were loaded with no end date. In the STIX format an expiry field is optional, which means that without one the entry stays valid indefinitely. Attackers abandon addresses fast, so stale entries mostly flag innocent traffic.

Fix: choose lifetimes by type and enforce them yourself. The Atlas’s sample policy sets a fortnight for network addresses, a month for web addresses, two months for domains and a year for file fingerprints. Treat the numbers as a starting draft and test them against how quickly your own alerts go stale.

Symptom: a sweep for known bad items came back empty, so we declared ourselves clear

Likely cause: confusing absence of a known item with absence of an intruder. In early 2024 several American agencies jointly warned about a state-backed group whose access in some networks had lasted five years or more, and the Atlas notes that the warning said conventional indicators were generally lacking.

Fix: pair every indicator sweep with a behaviour-based hunt. Choose one or two techniques, form a hypothesis, check you have the logs, and finish with a written result. In aircraft terms, a clean tick on the checklist does not replace listening to the engine. Our own notes on adding a glass-cockpit twin to the fleet make a similar point about new equipment: new instruments change what you must look for.

Symptom: hunts happen, but nobody ever acts on them

Likely cause: no hand-off. A hunt that ends in a chat message disappears. The Atlas wants every hunt to finish with a tangible output, whether a request for a new alarm, a log-coverage ticket, a handover to responders, or a record of a confident negative.

Fix: adopt the hand-off rule it describes: once a finding is validated, it goes to the owners of live alarm rules as a formal request. Track the share of hunts that produce something concrete; the Atlas’s worked quarter shows five of twelve, or 42 percent. Also test repeatability by asking a colleague who sat out the hunt to run it again.

Symptom: analysts disagree about what a report means

Likely cause: loose language. One person’s “probably” is another’s “maybe”. The Atlas recommends the NATO estimative words, where “likely” means 60 to 90 percent, with confidence stated separately and never in the same sentence.

Fix: adopt the word list as a house style, and for contested judgements use the two structured methods the Atlas names: an assumptions check and a competing-hypotheses table. It takes ten minutes and prevents a long argument.

Symptom: the board asks for numbers and we hand over counts

Likely cause: reporting activity rather than outcome. The Atlas dismisses raw tallies of items loaded, documents issued, hunts begun and coloured coverage charts as vanity measures.

Fix: report four ratios: how many priority questions have an answer, the portion of intelligence that triggered action, the portion of hunts that ended in a concrete output, and the portion of known gaps closed. Its sample quarter reads 83, 35, 42 and 44 percent. A low number is useful, because it points at the weakest link, much as a simulator debrief points at the weakest manoeuvre; compare the approach with the debrief habits in the note on adding an instrument-procedure simulator.

Symptom: we do not know where to begin

Likely cause: trying to build everything at once.

Fix: follow the Atlas’s decision tree. Without ninety days of usable log data, repair the data first. Without anyone to own findings, buy hunting as a service. Without stakeholders who ask questions, keep a lean advisory model. Otherwise build, beginning in month one with a named owner, a handful of questions, a graded list of sources, expiry dates and a seat in a sharing community.

A quick triage order

  1. Stop any automatic blocking that has not been measured.
  2. Put an expiry on every indicator you hold.
  3. Write the priority questions and match feeds to them.
  4. Run one small hunt and finish it with an artefact.
  5. Report ratios, not volumes, at the next review.

None of these steps needs new software. They need an owner, an hour a week and the discipline to write results down.

© 2024 ADVANCED INTERNATIONAL AVIATION ACADEMY. ALL RIGHTS RESERVED.

STAY CONNECTED WITH US: