Overview
The framework
This is the model itself: what a detection program has to do, in what order, and how you can tell a step is done without taking anyone's word for it. If you only read one part of the site, read this one. The rest is detail, and I say that as the person who wrote the detail.
On this page
The lifecycle
- Backlog line1. IntakeDone when: 6 scores, each with an evidence line, and 1 computed score
- Backlog line2. HypothesisDone when: 1 hypothesis sentence, 1 named log source, 1 named blind spot
- Backlog line3. ResearchDone when: 1 emulation run, fixtures saved, every field checked against a real event, 7 days of baseline counts, lab torn down
- Build line4. BuildDone when: lint passes, 1 conversion per target, 1 event it must match, 1 event it must ignore
- Build line5. ValidateDone when: fixtures pass on every replay engine, every converted query approved (matching its approved copy on a re-validation), 1 approval by someone other than the author (one-person team: the author after 24 hours)
- Live line6. DeployDone when: 7 days live with alerting off, alert volume at or under its threshold, then alerting on
- Live line7. TuneDone when: precision checked every 30 days, running query matches the approved copy at each check
- Live line8. RetireDone when: status deprecated, alerting off, record kept, retirement note written, replacement linked if one exists
In this part
AssumptionsExplains the operational realities behind the framework and the mechanics that follow from each.Read this to see if the framework is the right fit for your program.Sizing the practiceOutlines the core practices, the simplest version of each, and how role ownership scales with your team's size.Read this when deciding which parts of the framework to adopt first.First stepsUses your current log inventory, incident history, and available hours to choose your first 10 detections.Read this when you are starting a detection program from scratch.The lifecycleDetails each stage of a detection's life, the definition of done required to move forward, and the feedback loops that send work back.Read this when setting up your ticket workflow.
The practices
Each one works on its own, so take the ones that fix a problem you have and leave the rest on the shelf.
Detection zeroThe scored backlogCountable definitions of doneThe detection recordSecond-engineer reviewRetirement triggers and the precision floorThe metric setDetection as codeThreat-model intake
The other overviews: Build detections · Run and measure · The tools.