Skill verification glossary.

An assessment that adjusts the difficulty of each question to how you answered the previous ones. Because the questions follow your level, it can place you more precisely with fewer questions than a fixed test, and guessing is less useful. Lemma's skills tests are adaptive assessments; the method behind them is computerised adaptive testing.

See also Computerised adaptive testing

Lemma's on-demand tutor for explanations and practice drills between live sessions. Passed drills add a small, capped amount to the practice part of SCI; they supplement evidence from people and assessments rather than replace it.

A record where new entries are added and existing ones are never overwritten. On Lemma, corrections, revocations and disputes are recorded as new entries, so the history shows how a record changed and why.

The single identifier Lemma uses for a skill. Different spellings, such as "ReactJS" and "React.js", resolve to the same key, so evidence for one skill is never split across several names.

Computerised adaptive testing (CAT, also spelled computerized) is the method behind an adaptive assessment: software picks each question based on how you answered the previous ones, so a correct answer leads to a harder question and a wrong one to an easier one. Lemma's assessments use it on a 1 to 10 difficulty scale, asking 15 to 25 questions and stopping once your level holds steady.

See also Adaptive assessment

A report that a session was recorded incorrectly, for example marked complete when it did not take place. A reviewer looks at the record and either upholds it or voids the session; the decision and any changes it causes are added to the evidence history.

Anything that shows a skill in use and can be traced to where it came from: a completed live session, an assessment result, a linked project, or a confirmation from someone who saw the work. Evidence is what SCI is computed from.

Work that exists outside Lemma, such as a code repository, a design file or a deployed project, linked to a skill and checked against its source. It counts towards the proof part of SCI, up to a limit.

The part of SCI (up to 15 points) that reflects how recent your evidence is. Regular recent practice keeps it high; as evidence ages, it counts for less here, so a score describes current ability rather than past activity.

The part of SCI that looks for gaming patterns, such as most sessions being with the same partner or a small group trading confirmations. For normal use it is zero and has no effect; it can only lower a score, never raise it.

The credits used to book live sessions on Lemma. Teaching a session earns credits, and every movement is recorded in a ledger rather than as an edited balance.

A scheduled one-to-one video session on a specific skill, where one person teaches and the other learns. A session completes automatically when both people join, and the rating from the person you worked with feeds the practice part of SCI.

How Lemma suggests session partners: by which skills each person can teach and wants to learn, time-zone overlap, shared language and how reliably each person turns up to sessions.

A confirmation from another person that you have a skill. Only a small number of unique people count towards SCI, and reciprocal patterns are checked by the integrity check, so a large network cannot inflate a score.

The part of SCI (up to 25 points) built from completed live sessions, weighted by the rating from the person you worked with. Teaching counts for more than learning, and AI drills add a small, capped amount.

The largest part of SCI (up to 40 points): assessment results, external proofs and peer confirmations. Each source has its own limit, so no single kind of proof can fill it alone.

The dated list of evidence behind a skill, shown on the Skills Passport. It lets anyone reading a score see which sessions, assessments and projects it rests on.

The part of SCI (up to 20 points) that reflects whether you turn up. Attended sessions build it; no-shows and late cancellations reduce it.

The version of the rules used to turn evidence into SCI scores. Each Passport records the version that produced its scores, so a score can always be read against the rules in force when it was computed.

Lemma's score from 0 to 100 for a single skill, computed only from evidence. It adds four capped parts (practice, proof, reliability and freshness) and applies an integrity check that can only subtract. Every point traces back to specific evidence.

An exchange where two people teach each other skills instead of paying a tutor. On Lemma, a skill swap runs as live one-to-one video sessions: you teach what you know, learn what you do not, and pay with Lemma Credits that teaching earns. Each completed session is recorded as evidence for the skill it covered.

Producing evidence that someone can do what they claim, and recording where that evidence came from so another person can inspect it. It differs from a single test or certificate because it combines several kinds of evidence.

Hiring on evidence of what a candidate can do, rather than on degrees, job titles or years of experience used as proxies for skill. In practice it means a consistent skill check, such as an adaptive assessment or a work sample test, read alongside the work a candidate has already done.

Your portable record of skills on Lemma: an SCI score per skill with the evidence behind it. The Passport is a signed document, so whoever receives it can check that Lemma issued it and that it has not changed since; you choose how and with whom it is shared. The term is also used for government and industry schemes, such as construction site cards or national qualification records. Lemma's Skills Passport is separate from those and covers technical and product skills.

A live session where you are the one teaching. Teaching earns Lemma Credits and counts for more than learning in the practice part of SCI, because explaining a skill well takes a deeper grasp of it.

A digital credential in the W3C Verifiable Credentials format: a set of claims about a person, digitally signed by the issuer so anyone can check who issued it and that it has not been altered. Lemma's Passport document is also signed, but it is not issued in the W3C verifiable credential format.

Lemma's interface for organisations to check a shared Passport, invite candidates to assessments and receive results in their own systems. It is currently in early access.

A hiring test that asks a candidate to do a small piece of the real job, such as fixing a bug or critiquing a design, instead of answering questions about it. On Lemma the closest equivalents are live sessions, where a practitioner watches you work, and connected work from GitHub, Figma or Vercel.