The prior effect-first website-layer claim is retired.

This repo is no longer positioning a public bootstrap website or any shared guidance bundle as an efficacy mechanism.

The active reset is smaller:

• keep the six runnable examples as code-on-disk references
• test one local-only hypothesis directly
• make no public efficacy claim until that smaller experiment is actually completed

• / — archive/reset notice only
• /health — liveness check

The old routes (/bootstrap.txt, /adopt, /evals, /evals.json) are retired.

Start with:

1. examples/
2. experiment/PROTOCOL.md
3. experiment/results.json

The first restart experiment is:

• task: todo-repo-local-only-001
• arms: baseline, local-only
• trials: 3
• metric: objective compile + test pass/fail
• strict rule: local-only must beat baseline on all 3 trials or the restart is kill

npm install
npm run build
npm test
npm run experiment:report

• The previous website-layer claim was not established.
• The current local-only restart is defined but not yet run.
• There is no current public efficacy claim.