studylog-agent

PLATO Study Partner Agent — Research Study Partner for studylog.ai

Tracks student/researcher learning progression. Every lesson is logged to PLATO as a tile. Later agents can "don the shell" and continue from prior sessions seamlessly. Self-improving curriculum via agent learning from past sessions.

Overview

The studylog-agent is an agent that:

• Logs every study session to PLATO as a tile (like git commits)
• Presents material through PLATO rooms, tracking progress
• Don the Shell: later agents load past session context to continue curriculum
• Self-improving: curriculum knowledge accumulates in the PLATO vessel

Architecture

• PLATO room: studylog-ai
• Each study session → PLATO tile with:
  • Topic studied
  • Concepts covered
  • Questions asked
  • Understanding level (novice → apprentice → journeyman → master)
  • Session notes
  • Next steps
  • Resources used
• Vessel: curriculum knowledge accumulates over time

Quick Start

# Log a study session
python3 studylog_agent.py

# Don the shell — resume where you left off
python3 studylog_agent.py --resume

# Show study history
python3 studylog_agent.py --history

# Suggest next lesson
python3 studylog_agent.py --suggest

# Show concept mastery stats
python3 studylog_agent.py --stats

Tile Schema

Each PLATO tile captures a complete study session:

Field	Type	Description
topic	string	What was studied
concepts_covered	list[string]	Concepts learned this session
questions_asked	list[string]	Student questions during session
understanding_level	string	novice/apprentice/journeyman/master
session_notes	string	Agent observations
next_steps	list[string]	Recommended follow-up topics
resources_used	list[string]	Books, papers, tools referenced

Don the Shell

When a student returns, the agent:

1. Reads prior lesson tiles from studylog-ai room
2. Identifies where they left off (last topic, concepts, questions)
3. Resumes curriculum from that exact point

This is analogous to a sailor checking the ship's log before taking the helm.

PLATO API

The agent communicates with the PLATO Room Server at http://localhost:8847:

from studylog_agent import StudyLogAgent

agent = StudyLogAgent(student_id="alice")

# Log a session
agent.submit_session(
    topic="Constraint Theory",
    concepts=["hard constraints", "soft constraints", "penalty methods"],
    questions=["How do I express an OR relation as a constraint?", "What's the complexity of CSP?"],
    understanding="apprentice",
    notes="Struggled with penalty method formulation but got it after examples.",
    next_steps=["Lagrange multipliers", "Dual decomposition"],
    resources=["Dechter - Constraint Processing", "tscrump YouTube"],
)

# Resume
ctx = agent.don_the_shell()
print(ctx["next_suggestion"])

Files

studylog-agent/
├── studylog_agent.py     # Main agent
├── README.md             # This file
└── LICENSE               # MIT

License

MIT

Fleet Context

This agent is part of the Cocapn Fleet — a system of domain-specific PLATO agents that track, learn, and persist knowledge across sessions. Each agent writes to its own PLATO room, creating a distributed knowledge graph that grows over time.

Links

• Live site: https://studylog.ai
• Fleet: https://cocapn.com
• PLATO: https://github.com/SuperInstance/plato

Related

• studylog.ai — Live site
• studylog-ai-pages — GitHub Pages source