feat: polish final Learning Console and GitHub landing

[GwriPennar]

Base

• Base SHA: 611a5248a8eb20c5ec22946e111fac1229d77782

Exact files changed

• README.md
• apps/learning_console/app.py
• apps/learning_console/renderers/compact_experiment.py
• apps/learning_console/renderers/guided_experiment.py
• apps/learning_console/renderers/hardware_replication.py
• apps/learning_console/renderers/lesson_renderer.py
• docs/build-week/DEMO-SCRIPT.md
• docs/build-week/JUDGING-GUIDE.md
• docs/build-week/SUBMISSION-CHECKLIST.md
• src/quantum_folk_lab/hardware_replication_evidence.py
• tests/learning/test_build_week_console.py
• tests/learning/test_hardware_replication_evidence.py

Revised learner journey

1. Start with the Guided Experiment and reveal all 256 exact assignments.
2. Compare the registered bounded quantum method with known truth.
3. Move to real folk-family data and distinguish the first IBM validation from the replicated hardware landscape.
4. Use Foundations and Glossary for progressive support.

The opening and README now share the exact-first pitch and the Guided Experiment is the default nested tab.

Hardware chart architecture gate

Passed. Two Vega-Lite scatter plots use the existing validated evidence loader and add no runtime dependency:

• EXP-010D ideal versus hardware: 25 points from experiments/EXP-010D-hardware-parameter-landscape-run/hardware-analysis.json.
• EXP-010D versus EXP-011: 25 matched points from that file and experiments/EXP-011-dense-hardware-landscape-run/hardware-analysis.json.

Tests compare every plotted value directly with the governed JSON, enforce point counts, and fail closed on malformed cell evidence. Headline rho values also come from the governed read model.

Safety and scientific boundary

• No experiment result or registered evidence changed or regenerated.
• No IBM service was contacted.
• No credential was accessed.
• No QPU job was submitted.
• EXP-012 was not started.
• No AI behavior, audio, routing framework, plotting dependency, or broad refactor was added.
• Exact classical evaluation remains authoritative.
• No quantum-advantage, speedup, scale, generalisation, musical-truth, or commercial-superiority claim is made.

Validation

• focused Learning Console, Foundations, and hardware tests: 20 passed before the final chart-integrity assertion;
• full pytest: 252 passed;
• focused final hardware evidence tests: 5 passed;
• Ruff lint and formatting: passed;
• strict mypy for the extended governed loader: passed;
• release verification: passed;
• public-safety scan: passed;
• diff and experiment-evidence integrity checks: passed.

Human review required

This PR remains draft. A human visual review is required before merge, including normal-laptop first impression, chart readability, tab labels, README flow, and capture/approval of a real 256 Reveal screenshot if desired. No screenshot is committed in this PR.