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.