For agents

Every lesson is also written for machines. Your coding agent can learn the same practices you do, or teach you the course right in your terminal.

1. Install the course as skills

Works with Claude Code, Codex, Cursor, OpenCode and any agent that supports Agent Skills.

npx skills add Haple/system-one

In Claude Code you can also add it as a plugin marketplace:

/plugin marketplace add Haple/system-one
/plugin install system-one@system-one

This installs three skills: decision-models, the practices from the course that your agent applies whenever it writes a decision-model call (typed questions, one request, thresholds from costs, select don't generate, code for arithmetic); system-one-tutor, which teaches you the course interactively (try teach me lesson 1 of system one); and decision-audit, which finds prompt-and-parse LLM calls in your codebase that should be decisions, and reviews existing decision-model integrations (try audit our LLM decisions).

2. Read the course directly

• /llms.txt: index of every lesson in Markdown
• /llms-full.txt: the entire course in one file
• /lessons/<slug>.md: one lesson, with transcript, takeaways, practices and sources
• /api/course.json: structured course data

3. Paste this into any agent

Read https://learn-system-one.pages.dev/llms.txt and teach me the System One course on decision
models, one lesson at a time. After each lesson, quiz me and help me
apply it to the project in this directory.