autodidact

Autonomously plan, build, and reflect toward research goals within budget constraints.

Updated Aug 23, 2026

Search Tags:#automation #planning #research #ai #self-learning #autodidact

Ready to run

npx skills add https://github.com/leonardofhy/openclaw-workspace --skill autodidact

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autodidact
Source: https://github.com/leonardofhy/openclaw-workspace/tree/main/skills/autodidact
Command: npx skills add https://github.com/leonardofhy/openclaw-workspace --skill autodidact

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, torch, transformers, librosa, matplotlib, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Autonomous self-directed learning and research capabilities reduce manual cycles by autonomously planning, learning, building, and reflecting toward long-term research goals.

Core Features & Use Cases

Phase-aware bootstrapping and cycle orchestration (boot -> orient -> decide -> act -> record -> update) within bounded budgets.
Memory-driven planning and queue management that reads memory/state to select feasible tasks across explore, converge, and execute phases.
Reproducible, auditable learning loop: every cycle outputs events to memory and updates active.json to track budgets and progress.
Use case: An ongoing research program uses Autodidact to autonomously discover ideas, run deterministic baselines, and capture insights while Leo oversees direction.

Quick Start

Trigger the cron or run the bootstrap script to start an autonomous learning cycle immediately.

Frequently Asked Questions about autodidact

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate long-horizon AI research and self-learning cycles?▼

Autodidact automates long-horizon AI research by autonomously planning, learning, building, and reflecting toward goals. It operates in explore, converge, and execute phases within bounded budgets. Triggering the bootstrap script starts an autonomous learning cycle immediately.

What is phase-aware bootstrapping in autonomous research agents?▼

Phase-aware bootstrapping in autonomous research agents orchestrates cycles across boot, orient, decide, act, record, and update steps. It reads memory state to select feasible tasks and updates active.json to track budgets and progress.

How do I track budget constraints and progress in autonomous research workflows?▼

You track budget constraints and progress in autonomous research workflows by enforcing phase limits and recording cycle events to memory. The agent updates active.json logs with each event, allowing you to audit progress and adjust direction.

Do I need PyTorch and Transformers to run autonomous self-learning workflows?▼

Yes, you need PyTorch and Transformers to run autonomous self-learning workflows. The skill depends on numpy, torch, transformers, librosa, and matplotlib to support its advanced research, planning, and prototyping capabilities.

Can I use autonomous learning agents for building prototypes and documenting findings?▼

Yes, you can use autonomous learning agents for building prototypes and documenting findings. The agent discovers ideas, runs deterministic baselines, and captures insights within memory state and knowledge graphs while you oversee direction.

What are the limitations of using a memory-driven planning approach for AI research?▼

A limitation of memory-driven planning for AI research is the reliance on bounded budgets and phase constraints. You must oversee direction and audit logged events, as the agent autonomously selects feasible tasks from memory state without manual intervention.