Personal Study Agent

A chatbot answers what you ask. An agent has goals — yours — and works toward them across multiple conversations. Your study agent should know your test dates, your weak spots, what you studied yesterday, and what you should study today. That requires memory and planning, not just clever prompts.

Three things a study agent needs

1. 1Persistent memory: a doc the agent reads at the start of every session — your subjects, exam dates, recent topics, current performance
2. 2A planning step: at the start of each session it asks 'given your goals and yesterday's progress, what should we work on today?'
3. 3A reflection step: at the end of each session it updates the memory doc with what you covered and how it went
4. 4A tool: a way to access your past notes, syllabus, or practice problems
5. 5An adaptation rule: if you missed yesterday's session, it doesn't pretend you didn't

The starter prompt

Adaptation, not guilt

The agent should never shame you for falling behind. It should ask what changed, then re-plan. A guilt-driven study system fails within 2 weeks. An adaptive one survives finals week, sickness, and that one teacher's surprise project.

Applied exercise: 30-minute setup

1. 1Create a memory doc with: subjects, test dates, weak spots, last 3 study sessions.
2. 2Set up a Claude Project or Custom GPT with the system prompt above.
3. 3Run your first session — let it propose a plan.
4. 4At the end, ask it to update the memory doc.
5. 5Use it daily for one week, then iterate the system prompt.

The big idea: a study agent isn't a smarter chatbot — it's a system with memory, goals, and a plan that survives your bad weeks.