القائمة
Extract durable things the parent has said in chat and keep memory/preferences.md current — a Pulse-only skill, not triggered by a direct parent request.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Build a designed, self-contained HTML academic map of the child's subjects, topics, and materials, plus real coaching notes for the parent on how to teach them better.
Build a designed, self-contained HTML progress report for the child — how she's actually doing, and what to work on next — surfaced in the Progress tab for both parent and child.
Create clear, child-ready study material (notes, worked examples, a revision sheet) for a topic, matched to the child's level and their own uploaded materials, packaged as one activity.
Create a practice test for the child from their materials and progress — child-facing questions with a separate parent-only answer key, packaged as one activity.
Extract genuine interest signals from the child's own conversations and keep memory/interests.md current — a Pulse-only skill, not triggered by a direct parent request.
A fun, off-syllabus "something new to explore" activity for the child — a light cover page plus a chat-driven discovery, tailored to grade and interests learned over time. Parent-initiated, handed off like any other activity.
| name | update-preferences |
| description | Extract durable things the parent has said in chat and keep memory/preferences.md current — a Pulse-only skill, not triggered by a direct parent request. |
This is run by Pulse (the periodic check-in), never by a live parent request — the parent never asks for this directly, so don't mention it to them as a step you're doing; just fold the result into your normal Pulse reply if it's worth a line.
memory/preferences.md is a small, LIVING file — like the academic map and progress
report, it is fully rewritten each time, never appended to. It exists to remember
things the parent has told you in chat that aren't already captured by a dedicated
field (set_child_profile, set_parent_label) — so they never have to repeat
themselves in a later conversation. This includes general teaching preferences
("she does better with visual examples", "keep sessions short") that aren't specific
to one activity — a specific activity's own teaching_mode/hints_before_answer/
persona (set via create_learning_activity's own intake) always takes precedence
for that activity; this file is the standing default/context, not a per-activity override.
Read what's already captured: cat memory/preferences.md if it exists (it may
not, on a fresh family — that's fine, start from empty).
Scan for anything durable the parent has said recently — you already carry this
conversation's own context, so don't cat the whole of conversations/parent.json;
it's the one ongoing parent thread (web, WhatsApp, and Pulse all share it) and only
grows, so a full read gets more wasteful every time you run this. tail -c 20000 conversations/parent.json is plenty to catch anything since your last check; widen
it only if you have a real reason to think something durable is further back than
that AND still uncaptured. Look for things like:
Write memory/preferences.md as a short plain-Markdown bullet list (no HTML,
this file is never shown directly to the parent or child, it's Quill's own working
memory). Merge new durable statements in, drop anything clearly superseded or no
longer true (the parent said something different since), and keep the whole file
compact — a handful of bullets, not a growing log. If nothing durable has been said
since you last checked, leave the file as-is; don't rewrite it just to touch it.
Nothing else needs to happen here — this file is read at the start of parent and child conversations (see the system prompt) so future turns apply it automatically. You do not need to announce this update to the parent unless it's genuinely worth one line in your normal Pulse reply.
