Forsy-AI/agent-apprenticeship-seed-dataset

Agent Apprenticeship Seed Dataset

The living ecosystem where AI agents run automated workflow loops on any task, improve through execution, and turn each run into reusable work experience + data to improve future agents.

As agents move into long-horizon, economically valuable work, Agent Apprenticeship creates the open infrastructure where real-world tasks generate reusable learning signals and complex workflows advance through agent loops that turn execution into shared improvement.

Contents

The first seed dataset includes:

• 500+ curated seed tasks sourced and grounded from real world
• 495 reusable agent lessons
• 1000+ full agent execution traces
• 1000+ agent work episodes / task rollouts

The seed dataset spans specialized economically valuable tasks across domains and forms the first layer of the Agent Apprenticeship ecosystem.

Agent Apprenticeship is now available for anyone to start using with local agents including Codex, Cursor, Claude Code, OpenClaw, OpenCode, Hermes Agent, and custom agents, alongside different model providers. Users can run automated agent workflow loops locally, contribute agent learning signals back to the ecosystem, and use shared ecosystem signals to improve their own agents.

Dataset Structure

seed_dataset/
  traces/
  attempts/
  evaluation/
  learning_signals/
  tasks/

Use with Agent Apprenticeship

Install and initialize the CLI:

npx agent-apprenticeship init

Explore ecosystem experience:

apprentice ecosystem list
apprentice ecosystem search cloud
apprentice ecosystem inspect aa-seed-task-501
apprentice ecosystem pull aa-seed-task-501

Create an Experience Pack from ecosystem experience:

apprentice learn create aa-seed-task-501
apprentice learn preview <pack_id>
apprentice learn replay <pack_id>
apprentice learn keep <pack_id>
apprentice run "..." --experience-pack <pack_id>

GitHub ecosystem repo: https://github.com/Forsy-AI/agent-apprenticeship