#356356th of 389 in Skills, plugins and prompts
agent-apprenticeship
CLI that loops local coding agents with mentor models and saves reusable lessons
- Stars
- 1.6k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Jul 2026
- Language
- Python
0 on 16 Jun 20261,603 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
An npm CLI (`apprentice`) that runs a task through a local agent such as Codex, Cursor, Claude Code, OpenCode or a custom command, with a mentor model evaluating each loop iteration. Each run produces an Experience Compilation that can be installed as Runtime Training for later runs or exported to a shared ecosystem. A seed dataset of 505 compilations is published on Hugging Face.
Who it is for: Developers running local coding agents who want reusable task experience
Strengths
- Works with Codex, Cursor, Claude Code, OpenClaw, OpenCode, Hermes Agent or a custom command template
- Mentor providers include OpenAI, Anthropic, Gemini and OpenRouter, with keys stored locally
- Private Internal Only mode keeps experience from being contributed to the public ecosystem
- Publishes a seed dataset: 500+ tasks, 1000+ traces, 505 experience compilations
Weaknesses
- Needs a separate agent CLI installed plus a paid mentor provider API key
- Benefit of installed Runtime Training on agent output is not measured in the README
- Economic value estimates per task are claimed without a described method
- Requires Node.js tooling though the repository is detected as Python
What it needs
- Needs Node.js/npm, A supported agent CLI (Codex, Cursor, Claude Code, etc.), Mentor provider API key
- Models: OpenAI, Anthropic, Gemini, OpenRouter
Quick answers
Is agent-apprenticeship open source?
Yes. Its code is on GitHub under the MIT license.
Also in Skills, plugins and prompts
See all 389| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |