# ray-r-ren/agent-apprenticeship

The living ecosystem where AI agents complete tasks through workflow loops, improve through iterative execution, are evaluated by mentor agents or humans in the loop, and turn completed work into reusable work experience and data to improve future agents.

Repository: https://github.com/ray-r-ren/agent-apprenticeship
Canonical: https://ross.abutalabs.com/products/agent-apprenticeship
Homepage: https://forsy.ai
Language: Python
License: MIT
License Family: permissive
Topics: agent-economy, agent-experience, agent-learning, agent-traces, agentic-ai, ai-agents, autonomous-agents, claude-code, codex, cursor, hermes-agent, loop-engineering, openclaw, opencode, post-training, real-world-tasks, reinforcement-learning, agent-apprenticeship, ecosystem-learning, training-signals
Last push: 2026-07-06T15:17:36+00:00

## Health v2 (maintenance only)
Score: 63/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 91, release rhythm 59, longevity 5
- inputs: {"age_days": 75, "days_push": 58, "days_rel": 61, "gap_med": null, "n_releases_24m": 1}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1335, forks 59 (observed 2026-08-28T04:04:25.366585+00:00)

## What it is
Agent Apprenticeship is an open ecosystem and CLI tool where AI agents complete real-world tasks through iterative workflow loops, are evaluated by mentor agents or humans, and convert completed work into reusable experience data and training signals. It ships with a seed dataset of curated tasks, agent execution traces, lessons, and experience compilations to improve future agents.

## Use cases
- generate training data from real agent task executions
- run automated agent workflow loops locally with claude code or codex
- collect agent traces and lessons for post-training
- evaluate agent work with mentor agents or human review
- improve my agent's performance using shared ecosystem learning signals
- estimate economic value of agent tasks
- bootstrap agent fine-tuning datasets from real-world tasks

## When to choose
- you want to turn agent task executions into reusable training data
- you run coding agents like Claude Code, Codex, or Cursor and want iterative improvement loops
- you need a seed dataset of real-world agent tasks, traces, and lessons
- you want human- or mentor-in-the-loop evaluation of agent work

## When to avoid
- you need a simple single-shot agent framework without learning loops
- you require a hosted managed service rather than local CLI-driven workflows
- your use case is unrelated to agent training or experience accumulation

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, machine-learning, workflow-automation, data-generation, rag
- domain: machine-learning, artificial-intelligence, developer-tools
- platform: python, cross-platform, cli
- tags: agent-learning, agent-traces, post-training, loop-engineering, training-signals, mentor-agents, human-in-the-loop, agent-economy, seed-dataset, claude-code, codex, cursor, ai-agents, automation, nodejs

## Member repositories
- ray-r-ren/agent-apprenticeship (main) score 63

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:25.366585+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:44:24.673338+00:00, confidence not recorded.
  - readme: https://github.com/ray-r-ren/agent-apprenticeship (fetched 2026-08-28T04:04:25.366585+00:00, sha d8e3d72e67a1)
  - homepage: https://forsy.ai (fetched 2026-08-29T12:03:38.986862+00:00, sha f6f4c617e483)
- Data as of 2026-08-30T08:39:29.467469+00:00.