Improve AI workflows for learning path authoring and review

## Overview

This epic tracks work to improve the AI-assisted workflows around learning paths: creating them, reviewing them, and preflighting them before a PR is opened.

## Goals

- Strengthen the create / review / preflight workflows for learning paths
- Ensure these workflows live and are maintained properly in this repo, not just as Cursor-specific artifacts
- Keep the three workflows consistent with each other and with the current schema/authoring rules

## Scope

1. **Create learning path**: [`/create-learning-path`](.cursor/commands/create-learning-path.md), [docs/learning-path-authoring.md](docs/learning-path-authoring.md)
2. **Review learning path PR**: [`review-learning-path`](.cursor/skills/review-learning-path/SKILL.md) skill, [`/review-learning-path-pr`](.cursor/commands/review-learning-path-pr.md)
3. **Preflight learning path**: [`preflight-learning-path`](.cursor/skills/preflight-learning-path/SKILL.md) skill, [`/preflight-learning-path`](.cursor/commands/preflight-learning-path.md)

These originated as Cursor-authored skills/commands under `.cursor/`. This epic covers bringing them up to the same bar as the rest of the repo's AI workflows (accuracy, coverage, and native support in Claude Code) rather than just being present.

## Success criteria

- All three workflows are current with the latest schema and authoring rules
- A docs writer can run create → preflight → review end-to-end and get consistent, trustworthy results
- The workflows work equally well invoked from Claude Code and from Cursor

## Child issues

Link scoped work as sub-issues of this epic.