Tracking competency growth
Competencies exist to answer one question defensibly: what can this apprentice actually do, and what's the evidence? This page describes what the app actually computes today, not the fuller design it's still growing into — see what's coming for that.
A competency is a five-level ladder
Every competency a curriculum defines carries the same five named levels, each with its own description: Foundation, Apprentice, Practitioner, Journeyman, and Mastery Evidence.
{
"schemaVersion": 1,
"id": "git",
"title": "Git",
"description": "Git data model, commits, branching, merging...",
"levels": [
{ "level": 1, "name": "Foundation", "description": "..." },
{ "level": 2, "name": "Apprentice", "description": "..." },
{ "level": 3, "name": "Practitioner", "description": "..." },
{ "level": 4, "name": "Journeyman", "description": "..." },
{ "level": 5, "name": "Mastery Evidence", "description": "..." }
]
}What the app computes today
Only approved tasks and submissions count toward a competency — work still in progress, submitted-but-not-yet-reviewed, or sent back for revision contributes nothing. An apprentice cannot self-award progress, and neither can an unreviewed submission.
From there, today's rule is deliberately simple: for every task an apprentice has had approved, the app looks at which competencies that task declares and counts it once toward each. A competency reaches Apprentice once at least one approved task counts toward it, and Practitioner once at least three do. That's the whole rule — Foundation, Journeyman, and Mastery Evidence are written as descriptions of what real growth looks like, for a mentor to judge by eye, but the app doesn't assign them automatically yet.
This claim is computed fresh every time, not stored anywhere as its own record. There's no separate competency ledger file to fall out of sync — the claim is only ever a live read of whichever submissions and reviews are actually approved right now in this repository.
What you can do as a mentor, today
- •See which competencies a task targets, right on the task itself, before you review it.
- •Trust that every task counted toward Apprentice or Practitioner traces back to a submission you (or a prior mentor) actually approved — see Reviewing submissions.
- •Use a competency's own level descriptions as your own rubric when you want to say more than "Apprentice" or "Practitioner" — that judgment is currently yours to make and record in your review comments, not something the app tracks for you yet.
The guarantee
Every competency claim you can see traces back to approved evidence you can open and inspect. If it can't, it isn't real progress yet — it's a suggestion.
What's coming
A fuller competency engine is planned but not yet built. It's worth knowing the direction, so you don't mistake today's simple heuristic for the finished design:
- •A competency map you can open and drill into, with the full evidence list behind every level.
- •Assigning or changing a competency level yourself, with the change recorded durably in Git.
- •An AI-suggested level, always clearly labeled non-authoritative and requiring your approval before it means anything.
- •Gap detection and suggested next tasks, to steer what an apprentice should work on next.
None of this changes the guarantee above — it only adds more ways to see and act on evidence that's already there.