[astrapi69]

Summary

adaptive-learner-content-template#92 moved the content rules to the engine. Its wave carried only the template-owned files, so this repository's own README.md and docs/GETTING-STARTED.md still describe the old tooling. This PR updates the stale passages and nothing else.

• Quality rules. free_text and picture_choice no longer need distractors; that requirement was dropped by decision in learn-content-engine#185. The rules are now enforced by the engine gate, not the validator. "purpose": "bridge" and "purpose": "quiz" are explained.
• Validation failure example. The example message is a structure-gate failure, since that is what validate_content.py still reports.
• Dependencies. Where named, the validator's dependencies point to requirements.txt (it also needs regex), and Node.js/npm is listed for the engine gate and the generator.
• Generator. It checks every draft against the mirrored schema and then the pinned engine, and it stops when the engine is not installed. The README no longer describes an engine gate that runs "when installed, otherwise in CI" after generation.
• Engine gate before push. It also enforces the quality minimums.

Not every passage exists in every repository; each change applies only where the stale text was found.

Test plan

• Before the change, the stale phrases were found with git grep on origin/main of the template and the nine content repositories. After the change, a scan for the same phrases reports nothing.
• scripts/check_prose.py reports no new finding in the two changed files.

🤖 Generated with Claude Code