Add data fabrication anomaly assistant

[drsteinerdj]

/claim #16

Summary

• Adds a distinct data-fabrication-anomaly-assistant/ slice for the AI-Powered Research Assistant Suite.
• Reviews synthetic manuscript data packets before AI peer-review output is trusted.
• Flags repeated measurement rows, invalid collection timestamps, terminal-digit preference, over-smooth measurement sequences, and perfect between-group separation.
• Emits deterministic RELEASE / REVIEW / HOLD decisions plus JSON, Markdown, SVG, and MP4 reviewer artifacts.

Scope boundary

This is a data-forensics anomaly guard. It is separate from broad assistant suites, model-assumption diagnostics, preregistration consistency, reviewer calibration, citation recency, statistical power, missing-data sensitivity, Bayesian prior sensitivity, structured abstract checks, and assistant-output disclosure slices.

All data is synthetic. The module does not call external AI APIs, publisher APIs, private manuscripts, live datasets, credential stores, payment systems, or reviewer identity services.

Demo artifacts

• data-fabrication-anomaly-assistant/reports/demo.mp4
• data-fabrication-anomaly-assistant/reports/risky-review.json
• data-fabrication-anomaly-assistant/reports/risky-review.md
• data-fabrication-anomaly-assistant/reports/summary.svg

Validation

cd data-fabrication-anomaly-assistant
npm run check
npm test
npm run demo
npm run demo:video
ffprobe -v error -select_streams v:0 -show_entries stream=codec_name,width,height,avg_frame_rate -show_entries format=duration,size -of default=noprint_wrappers=1 reports/demo.mp4
cd ..
git diff --check

Observed locally:

data fabrication anomaly assistant tests passed
codec_name=h264
width=960
height=540
avg_frame_rate=12/1
duration=4.000000
size=23155