Build Week 2026.2: Verifiable Offline Evidence

Final Build Week evidence release.

Judge path

1. Download plotproof-review-evidence-build-week-2026.2.zip.
2. Extract it and open review_report.html in any modern browser. No MATLAB, API key, JavaScript, or network connection is required to inspect the report.
3. To reproduce the figures from source, follow the one-command MATLAB path in the repository README.
4. To verify copied artifacts, run python3 scripts/verify_review_bundle.py --manifest <bundle>/review_bundle_manifest.json --root <bundle>.

What this release demonstrates

• three real MATLAB R2025a candidate renders from bundled synthetic data
• GPT-5.6 review evidence validated before MATLAB repair
• semantic rejection of an invalid confidence-band interpretation
• controlled repair allowlist with no model-authored MATLAB execution
• embedded before/after comparison, scores, findings, and provenance
• 12 evidence artifacts recorded with byte counts and SHA-256 digests

Asset SHA-256: 6e4886124c849b112938db836b4849ee380e1700ed227c1cb98ecc3bd569a0d1

PR: #45
Main Quality run: https://github.com/Kkkakania/matlab-plotting-skill/actions/runs/29852278547

Build Week 2026.1: Adversarial contract evidence

This incremental Build Week release hardens the Codex-to-MATLAB boundary with a reproducible adversarial benchmark.

• Accepts one checked GPT-5.6 control review
• Rejects 14/14 hostile mutations fail closed
• Covers executable-action injection, candidate/model spoofing, field smuggling, invalid scores, contradictory verdicts, oversized findings, and malformed JSON
• Commits human-readable and machine-readable evidence
• Runs the benchmark in the release gate

Reproduce:

./scripts/run_review_contract_benchmark.py \
  --validator scripts/validate_plot_review.py \
  --manifest examples/review/multi_series_manifest.json \
  --review examples/review/multi_series_review.json \
  --out /tmp/plot-review-contract-benchmark

Verification: PR #43 passed both Quality checks, and main Quality run 29636319169 completed successfully. All inputs are synthetic and reports contain no private paths.

Build Week 2026: Evidence-backed plot review

This release adds a Codex-native scientific plot review loop to the MATLAB Plotting Skill.

• Render 2-5 ranked MATLAB candidates from the same data and goal
• Validate GPT-5.6 reviews across five quality dimensions
• Reject unknown candidates, models, fields, scores, and repair actions
• Apply only controlled figure-object repairs
• Export final PNG/SVG, before/after comparison, and Markdown/JSON evidence
• Reproduce the complete path with bundled synthetic data

Run:

MATLAB_BIN=/Applications/MATLAB_R2025a.app/bin/matlab ./scripts/run_review_demo.sh --out /tmp/matlab-plot-review-demo

See docs/build-week-2026.md for the pre-event/new-work boundary and PR #42 for the reviewed implementation.

v0.1.100

• Completed TASK-068-scatter_relationship-report.
• Added MATLAB regression coverage proving Markdown and JSON reports record scatter_relationship and its generated output.
• Release gate passed locally with MATLAB R2025a.

v0.1.99

• Completed TASK-067-scatter_relationship-vector-render.
• Added MATLAB regression coverage proving scatter_relationship.svg and scatter_relationship.pdf are generated and non-empty.
• Release gate passed locally with MATLAB R2025a.

v0.1.98

• Completed TASK-066-scatter_relationship-png-render.
• Added MATLAB regression coverage proving scatter_relationship.png is generated and non-empty.
• Release gate passed locally with MATLAB R2025a.

v0.1.97

• Completed TASK-065-scatter_relationship-explicit-cli.
• Added MATLAB regression coverage for explicit scatter_relationship output and reports.
• Release gate passed locally with MATLAB R2025a.

v0.1.96

• Completed TASK-064-scatter_relationship-selection-rule.
• Added MATLAB coverage proving simple paired numeric data selects scatter_relationship.
• Release gate passed locally with MATLAB R2025a.

v0.1.95

• Completed TASK-063-scatter_relationship-demo-data.
• Added MATLAB coverage proving bundled scatter demo data supports schema inference and x-y rendering.
• Release gate passed locally with MATLAB R2025a.

v0.1.94

• Completed TASK-062-scatter_relationship-data-contract.
• Documented paired numeric-column expectations and misuse cases for scatter_relationship.
• Release gate passed locally with MATLAB R2025a.