A daily maths programme for a single child, where the teaching decisions are made by Claude Code agents on the parent's own machine. The child sees a plain web page with no model behind it. Every session is written, measured and independently re-solved before they open it — and the number that decides whether they move on is not the number they are shown. Try a session →
Read each lane left to right. The top lane is the child, on their own device, with no AI in it. The bottom lane is the parent's machine, where one command runs the whole pipeline. The middle lane is the record both halves depend on. Dashed markers are the two points where the pipeline stops and waits for a person.
submitted/grade
The gate is conditional, not ceremonial: an independent re-solve that passes saves by itself. The pipeline stops for a person only when that re-solve fails, or when a second failed check forces a backtrack to a prerequisite — the two cases where a human must actually decide something.
The single most consequential design decision here, and the one that took a bug to find. A score that counts help as success is the exact number you must not use to decide what comes next.
How many questions ended correct, with any amount of help — a wrong try, three clues, a revealed answer. This is the morale number, large on their own page.
Right on the first try with zero clues and no reveal. This is the mastery number, and the only one progression reads.
Each lesson is split into cycles — teach a bit, then one guided question on that bit, with full support up. It measures the explanation, not the child.
Why a guided question is excluded from the score entirely. It sits immediately after the explanation that taught it, with every clue available. Counting a heavily scaffolded answer toward mastery would inflate precisely the number that decides whether the child moves on. So it gets its own record — concept slug, outcome, whether a second angle was needed — and that record moves the teaching, never the verdict.
Each is a separate agent definition with its own instructions and its own tool access. Read-only agents are read-only by configuration, not by being asked nicely — which is the whole point of the ones that judge.
Marks by working each answer itself, diagnoses why each wrong one was wrong, and decides the verdict and next stage.
Writes the child-facing card from the grader's diagnosis — steps first, then the rule.
Writes the lesson as cycles, each with a guided question and a pre-written second angle for a wrong answer.
Writes the questions: the ask, the answer, the clue ladder, and the wrong answers it expects.
Draws the diagrams the session requested, from a fixed vocabulary of shapes — never free SVG, never raw coordinates.
Re-solves every question cold, with no sight of how it was written. Fails on any disagreement — or any ambiguity.
Researches the published framework for a board and stage, and proposes topics in prerequisite order.
Reviews that draft blind, with no sight of how it was derived, for false and missing prerequisites.
There is no skip button, because every rung of this ladder ends in an answer. The cost is that someone has to predict the wrong answers in advance — which is work the authoring agent does, question by question.
| Wrong try | What the child gets | What the record gets |
|---|---|---|
| First | The coach names that specific mistake, if it was one of the predicted ones. The clues unlock; none opens yet. | The slip code |
| Asks for a hand | The next unseen clue opens and is read aloud. | Clues used |
| Second and after | The next clue opens by itself — the child never has to ask to stop being stuck. | Clues used |
| Clues exhausted | Show me how — the answer, with the last clue as the method, and "good on you for keeping at it". | revealed |
| On a check | No clues at all, and no answer ever shown. After four tries, a Move on button, and nothing else. | exhausted |
Those last two rows are different diagnoses, and the system treats them that way. revealed means the clues were not enough. exhausted means the child could not get there with no help available and no pressure to perform — the strongest possible signal to re-teach, and two of them on one topic forces a backtrack to its prerequisite. Both outrank a plain wrong answer. Neither is ever described to the child as failure.
No agent writes SVG. A diagram is a claim about what must be true, drawn by one agent and measured back by a script — because a wrong picture teaches the error exactly as directly as a wrong answer key does.
The writing agent states the facts: a triangle with sides 6, 6 and 9. It does not draw anything.
A separate agent builds the drawing from a fixed vocabulary of shapes — no coordinates invented by hand.
The checker re-measures the drawing. Sides that are not 6, 6 and 9 block the session from saving.
A question can declare words the picture must not print, and the checker reads the labels back to enforce it.
It also refuses a diagram whose labels overlap. A stimulus the child cannot read is as broken as one drawn wrongly — and nobody reviewing the JSON would ever have noticed.
The governance question in one table. The child's device holds a key that can do two things; the agents that judge can do nothing at all.
| Component | Runs where | Session record | Child's progress | The web |
|---|---|---|---|---|
| The child's page | Their device | Post answers only | No access | No model, no key |
| grader | Parent's Mac | Read | Write, via one script | No access |
| feedback | Parent's Mac | Read | No access | No access |
| teacher and author | Parent's Mac | Write a draft file | Read | No access |
| illustrator | Parent's Mac | Write drawings only | No access | No access |
| verifier | Parent's Mac | Read | No access | No access |
| curriculum | Parent's Mac | No access | Draft, on approval | Research |
| The checker script | Parent's Mac | Read | No access | No access |
| The parent | — | Write | Write · owns the topic queue | — |
Three things to notice. Nothing reaches the database except through a script that enforces the rules a prose instruction cannot — a passed check always sets mastered, a failed one never does, and a second failed check refuses to save at all without a named prerequisite to fall back to. No agent can queue a second session while one is pending. And the parent owns the topic queue outright: a topic they pause leaves the queue and keeps its entire history, and no agent may silently put it back.
About half an hour, most of it the database. The only paid part is Claude Code — the grading and lesson writing is the piece no script replaces.
Atlas's free tier, no card. npm run check tells you in plain English which of the four usual mistakes it is.
One for the child's device, which can only fetch and answer. One for the dashboard. Local dev skips both.
/setup-child researches the board's published framework and proposes topics in prerequisite order.
/next-session. There is nothing to grade yet, so this builds the opening session directly.
npm start. The child's page at /, yours at /dashboard.html.
Optional. Her built-in lines already ship as audio; this records her on your child's own sessions.
Optional, and only if the child needs it on their own device. Local-only works completely.
They do the session. You run /grade. That is the whole routine.
The scripts must run on your real machine. Cloud sandboxes are firewalled off from your database, which is the correct outcome: the one copy of a child's record lives somewhere you control.