Review queues: bugs, audio, exams, capstones & projects
What lands in each review queue, and what happens when you action it.
- Every kind of learner work has a queue — nothing waits in a shared inbox
- Actioning a queue item awards XP, records the review and notifies the learner automatically
- AI does the first pass (grades, transcripts, flags); you make the call
Five queues, one habit
Admin → Review — the daily queue
- Pending bug reports — decide valid / invalid / duplicate and rate quality. A valid bug awards XP scaled by severity and quality; the learner is notified with your feedback.
- AI grades flagged for audit — free-text answers where the AI was unsure land here for a human check.
- Reverse-engineer submissions — the AI's score is prefilled; you set the final score and feedback.
Admin → Audio — spoken work
Teach-backs, daily reflections, capstone pitches and code walkthroughs, each with playback, transcript, AI grade and an AI-suspected flag when the speech looks generated. Jump straight to the learner's work page from any item.
Admin → Exams — integrity & attempts
Every graded attempt with its behavioural signals — tab blur, tab hidden, fullscreen exit, paste, copy — rolled into a Low / Medium / High risk rating, riskiest first. Click Read to see the actual answers. Anti-cheat also includes randomized question order, timed auto-submit and a one-attempt entrance lock.
Admin → Capstones — score & publish
Submitted capstones (brief, repo, demo) await your score. Reviewing can publish the best work to the learner's public portfolio.
Admin → Projects — apprenticeship work
Author real incubated projects, assign apprentices and tasks, and set project status (open / active / done). Members and task counts are visible per project.
Every queue is fenced to your own organization; mentors see their cohort, admins see everything in the org.
The journey, step by step
| Step | What you do | What you get |
|---|---|---|
| Validate bugs | Admin → Review: decide valid / invalid / duplicate + quality | XP is awarded by severity and quality; the learner is notified with your feedback |
| Audit AI grades | Check answers the AI flagged for manual review | Uncertain machine grades get a human eye |
| Sign off submissions | Score reverse-engineer submissions (AI score prefilled) | The final mentor score replaces the AI's |
| Listen to audio | Admin → Audio: play teach-backs, reflections and pitches | Transcript, AI grade and AI-suspected flags help you judge fast |
| Check integrity | Admin → Exams: scan graded attempts by risk | Flag patterns (blur, paste, fullscreen exits) surface who to talk to |
| Review capstones & projects | Admin → Capstones / Projects | Scored capstones can be published to the learner's public portfolio; projects get members, tasks and status |
Frequently asked
What lands in the main Review Queue?
Three things: pending bug reports awaiting a validity verdict, exam answers the AI flagged for manual audit, and reverse-engineer submissions awaiting mentor sign-off.
What does marking a bug valid do?
It awards XP based on severity and the quality you pick (excellent / good / weak), records the review, and notifies the learner with your feedback and the points earned.
Are integrity flags proof of cheating?
No — treat them as review triggers. The exam surface logs tab blurs, paste/copy and fullscreen exits; look for patterns, then read the answers before judging.
Related guides
One page per learner: vitals, integrity, AI usage, every artifact they produced — and the levers to coach them.
ReadRegister the apps learners test, sync their schema and repo to unlock walkthroughs, and run the Bug Board.
ReadCreate flashcards, coding challenges, lessons and case studies; tune AI tutor personas and role-play scenarios.
Read