AI Productivity · AI-TEAM
Multi-Agent Orchestration and Naming
Defining roles, naming every agent, isolating their work, handing off through files, merging results safely
What will be measured
- Format
- Hands-on: terminal tasks. No trivia — you do real tasks.
- Domains
- Naming · Isolation · Hand-off · Merging
- Duration
- about 25 minutes
- Levels examinable
- L1–L4
- AI policy
- AI allowed (and expected)
- Scoring
- Correctness first; efficiency (time, keystrokes, steps vs an expert reference) adds a bounded bonus. Graders check outcomes, so any valid approach passes.
- Credential
- Module badge, valid 24 months. $29 per certification attempt; practice is free.
Level definitions
- L1Foundation. Runs one named agent
- L2Proficient. Between Foundation and Advanced
- L3Advanced. Two named agents in separate worktrees
- L4Expert. Three to four agents with distinct roles
- L5Master. Repeatable multi-agent workflow
Integrity and privacy — no surprises
- Standard mode (L1–L3 on your public certificate): email verification, focus/tab detection, keystroke timing telemetry, randomized parameterized tasks, server-side timer.
- Proctored mode (required for public L4–L5): government ID and selfie check, webcam snapshots and optional screen share, reviewed only if a session is flagged. Recordings are deleted after 90 days.
- Unproctored results are capped at L3 on the public certificate; you can book a proctored retake to confirm L4 or L5. Every certificate shows whether it was proctored.
- Retakes: wait 14 days per module, maximum 3 attempts per quarter. Your highest valid level shows on your profile.
- We never show raw keystrokes or video to employers. You own your results. Data notice
Start this exam
Create a free account to take exams and keep your results.
Sign up to start