Glossary

A precise reference for the vocabulary used across QualifyCode. Terms are listed alphabetically.

A

Active — An assessment status. The assessment is paid for and live. Candidates can open their invitation links and start.

Admin — A workspace role. Owns the workspace: members, billing, settings, and every assessment in the workspace.

Anti-cheat — The class of signals that surface potential inappropriate help during an assessment: paste, focus events, AI assist, and so on. Every signal is a lead for a reviewer, not an automated verdict.

Approval — The optional step that signs off an assessment's design (problem set and launch-controlled fields) before launch. Performed by an authorized recruiter or hiring manager, under a single or dual approver policy. See Assessment approval.

Assessment — A coding test. Created via the wizard, paid for, sent to candidates, and reviewed.

Attempt — A candidate's run of an assessment. A candidate may have more than one attempt in your workspace if they have taken multiple assessments.

Audit trail — The complete history of every change to a candidate's verdict, recorded as an event with actor, source, reason, and timestamp.

Automation — The platform's ability to apply the Role Bar to candidates. Runs in three modes — Off, Trial (records conclusions for comparison, affects nobody), and Active (decides) — with a kill switch and a reversals log. See Automation.

Awaiting payment — An assessment status. Submitted for payment, checkout open.

B

Bar — Short for Role Bar. Bulk bar — The action bar in the screening queue that appears when one or more candidates are selected. Apply Pass / Review / Reject in bulk, or open Compare.

Bulk paste — A signal that a large block of code arrived at once rather than being typed. Recorded with the timestamp.

C

Calendar tool — A timeline pin that marks a notable moment in a candidate's work.

Candidate — A person invited to take an assessment.

Candidate app — The web app candidates use to take a test. Separate from the dashboard. No account required.

Candidate status — Where a candidate is in their journey: PENDING, SENT, STARTED, COMPLETED, EXPIRED.

CETS — Code Evolved Telemetry Score. The analysis that combines code telemetry and run results into an integrity report and a recommendation.

Channel — A place a notification can be delivered: in-app, email, Slack, or Microsoft Teams. See Notifications.

CLEAR — A screening verdict. The candidate has been advanced by a reviewer.

Cohort — The set of candidates in the same workspace and same role, used to compute a percentile.

Completeness — Whether a candidate's evidence (score, CETS, video, integrity) is sufficient to decide.

Composite — The overall assessment score combining integrity and quality.

CRM ID — A free-text field on an assessment for your own tracking.

D

Dashboard — The web app recruiters and admins use to manage assessments, candidates, and results. Separate from the candidate app.

Decision — See verdict.

Decision Room — The single page that aggregates every attempt and every piece of evidence for a candidate.

Direct message — A private Slack or Microsoft Teams message sent to one person. You link your own account to receive direct messages; it is separate from the workspace's shared channel. See Notifications.

Disqualify — A flag disposition. The candidate is rejected on this signal alone. Not allowed for fraud signals.

Draft — An assessment status. Created, not yet paid for. The wizard autosaves drafts to the server.

Dossier — A candidate's record across assessments: history, notes, skills, attempts, shortlists.

E

Escalate — A flag disposition. The candidate is sent to a person for review.

Executive summary — The plain-language summary at the top of an integrity report.

Expired — A candidate or assessment status. The schedule window closed before completion.

F

Flag — An integrity signal. Eleven declared (llm_assisted, bulk_paste, transcription, resume_mismatch, phone_detected, multiple_faces, looking_away, tab_switch, fullscreen_exit, low_effort, no_show); eight emitted by the platform today.

Fallback verdict — The verdict that applies when the bar cannot settle. Fixed to REVIEW — ambiguity resolves to a human.

Fraud signal — A flag in the integrity-fraud set: LLM-assisted, bulk paste, transcription, phone detected, tab switch, multiple faces. The bar may never disqualify on these.

G

Gateway — A payment processor. Razorpay (India) and Stripe (all other regions).

Grade threshold — The score above which a number is rendered in success, above which in warning, and at or below which in destructive.

H

Handoff Pack — A focused interview brief for the next round.

Hiring manager — An approver designation used by the assessment approval step (dual approval requires one recruiter and one hiring manager). Separate from the admin and recruiter workspace roles.

Hold — A wallet entry. Credits reserved for an assessment in flight.

I

In Screening — A screening verdict. Evidence is still arriving.

Inform — A flag disposition. The flag is shown to the reviewer but does not change the outcome.

INSUFFICIENT_DATA — A suggested-verdict value. Used as a machine suggestion only, never a storable verdict.

Integrations — The Notifications tab where an admin connects Slack and Microsoft Teams for the workspace. See Notifications.

Integrity aggregate — The integrity component of the composite score.

Integrity guard — A safeguard on the bar: a hard rule that the bar refuses to act on certain signals.

Invitation link — A private URL the candidate uses to start the assessment. Tied to one candidate, one assessment.

J

Job ID — A record in the platform that ties a candidate to an assessment. Sometimes called a "candidate assessment".

K

Keyboard triage — The keyboard-first interaction model of the screening queue: j, k, c, r, x, Enter, Esc.

L

Ledger — The complete record of every credit that has entered or left the workspace wallet.

Logical leap — A signal that the code contains a sudden logical step inconsistent with the rest of the session.

M

Manifest — A short list of what the assessment can speak to and what it cannot. Stated as a list of limits on the Handoff Pack and the Decision Room.

Microsoft Teams — A chat integration for notifications, alongside Slack. A workspace admin connects it once; direct messages are linked per person. See Notifications.

My scope — The scope toggle on the assessment list, for admins. Limits the view to assessments the admin owns.

N

Natural evolution — A signal that measures how much the code developed naturally through normal editing.

NEW — A screening verdict. The candidate has not been decided. Surfaced as "Pending" in the UI.

Notification — A message about hiring activity (a candidate starts, submits, or completes; an analysis is ready; a payment is captured; the wallet is low). Delivered in-app, by email, to a shared Slack channel, or as a direct message. See Notifications.

Notes — A private comment you can add to a candidate's dossier.

O

Override — A reviewer's manual override of the CETS recommendation on a specific attempt. Recorded as a separate event from the screening verdict.

OWN — The implicit share permission held by the creator of an assessment and by any workspace admin. OWN can edit, share, and delete.

P

PENDING — A candidate status. Added to the assessment but the invitation has not been sent yet.

Percentile — Where a candidate's best score sits among same-organisation, same-role peers. null when the cohort is too small; the reason is shown in place of a number.

Policy — The Role Bar. The published rules that map a candidate's analysis into a verdict.

Pin — A timeline pin. A notable moment in a candidate's work, color-coded by type.

Preview impact — A read-only preview that shows how many existing candidates would change verdict under a proposed bar.

Private link — The candidate's invitation link.

Problem — A coding challenge from the shared library or an AI variation.

Q

Quality aggregate — The quality component of the composite score.

Queue — The Screening console. The list of candidates across your assessments, organized by verdict, with filters, sort, and bulk actions.

R

Recommendation — The machine's read on a candidate, in six buckets: STRONG_HIRE, HIRE, LEAN_HIRE, LEAN_NO_HIRE, NO_HIRE, STRONG_NO_HIRE.

Recruiter — A workspace role. Creates and edits their own assessments, sees their own candidates.

Reject — A screening verdict. The candidate has been ruled out.

RELEASE — A wallet entry. A hold was released (cancellation or completion).

Report — The integrity report. The page that brings together code, results, and video for one attempt.

Review — A screening verdict. The candidate has been flagged for a second look.

Role Bar — The policy. The published rules that map a candidate's analysis into a verdict.

Role title — A free-text field on an assessment. Used to group candidates into cohorts and to filter the screening queue.

S

Sample data — A dev-mode escape hatch that substitutes checked-in fixtures when a request fails. Force-disabled in production.

Schedule window — The start and end datetimes of an assessment, in the workspace timezone.

Scope — Where a policy applies: ORGANIZATION, ROLE, or ASSESSMENT.

Screening console — The Screening section of the product.

Screening verdict — A recorded decision about a candidate, in five stages: NEW, SCREENING, CLEAR, REVIEW, REJECT.

Session — A single run of a candidate through an assessment. Sometimes called an "attempt".

Share — Per-assessment access for a specific user. VIEW or EDIT.

Shortlist — A curated collection of candidates for a specific role or hiring round.

Slack — A chat integration for notifications. A workspace admin connects it once; the workspace can post to shared channels, and each person can link their own account for direct messages. See Notifications.

Smart variation — An AI-generated rewrite of a problem. Shared across the workspace, not unique per candidate.

Status dot — A small colored indicator used in pipeline bars and status banners.

Stripe — A payment gateway. The default for non-Indian regions.

Subprocessor — A third-party data processor. Listed in Data and privacy.

Suggested verdict — The machine's read on a candidate, in the same five stages as the screening verdict plus INSUFFICIENT_DATA.

T

Tag — A custom label on a candidate, used to filter the screening queue.

Template — A reusable blueprint for an assessment. Saved from an existing assessment, edited to a new version, and instantiated as a new assessment.

Template version — The version of the template an assessment was created from. Recorded on the assessment for traceability.

Timeline pin — A notable moment in a candidate's work. Click to seek the video and load the code snapshot.

Token — A wallet credit. One token = one candidate seat.

Transcription — A signal that typing cadence matched copying from another screen rather than composing.

U

Unarchive — The reverse of archive. Brings a template back into the active list.

Unshare — The reverse of share. Removes a user's access to a specific assessment.

V

Verdict — A screening verdict. See also recommendation (the machine's read) and override (a reviewer's manual override of the CETS recommendation).

Version — Each edit to a template or policy creates a new version. Past versions are preserved for traceability.

Video audit — The analysis that processes the recorded video in chunks and produces per-chunk attention and integrity signals.

Video chunk — A short piece of the recorded video, stored in object storage.

VIEW — A share permission. Read-only access to a specific assessment.

W

Wallet — The workspace's pre-funded pool of credits. Used to pay for assessments automatically.

Wallet entry — A record in the ledger. CREDIT, DEBIT, HOLD, or RELEASE.

Workspace — A company's shared space in QualifyCode. Users in the same workspace share the same assessments, candidates, and ledger, gated by role.

What to do next