Add research gap replication planner

[firstedition0123]

@algora-pbc /claim #16

Contributes to #16.

Summary

• Adds a self-contained research-gap-replication-planner module for the AI-Powered Research Assistant Suite bounty.
• Focuses on the gap-to-replication planning boundary rather than another broad assistant suite, evidence-grounding module, or protocol-trace module.
• Generates auto peer-review findings for unsupported claims, citation gaps, statistical risk, and domain-specific methodology checks.
• Scores reproducibility confidence from raw data, clean pipelines, environment locks, test sets, reported outputs, and previous attempts.
• Ranks under-replicated research intersections using unresolved/negative-result signals, replication count, corpus activity, lab fit, and previous reproducibility friction.
• Includes synthetic sample data, requirement mapping, Node tests, CLI demo, SVG frame, and a 4s MP4 demo.

Why this slice is distinct

Existing #16 submissions cover broad assistant suites, evidence grounding, protocol traces, quality gates, and readiness engines. This PR specifically targets the research-gap-to-replication decision point: deciding which unresolved corpus intersections a lab should replicate first, and what evidence is needed before execution.

Demo

• research-gap-replication-planner/docs/demo.mp4
• research-gap-replication-planner/docs/demo.svg
• cd research-gap-replication-planner && npm run demo

Demo output:

Planner status: hold
Top opportunity: alzheimer + crispr + single-cell
Top replication decision: prioritize
Top peer-review flag: Claim has no supporting evidence artifact.
Weak reproducibility project: circadian-metabolomics-2026
Audit hash: de1b1fa44644dc24...

Verification

cd research-gap-replication-planner
npm run check
ffprobe -v error -select_streams v:0 -show_entries stream=codec_name,width,height,duration -of default=noprint_wrappers=1 docs/demo.mp4
cd ..
git diff --check
rg -n 'api[_-]?key|private[_-]?key|password|secret|token|wallet|ssn' research-gap-replication-planner

Local results:

research-gap-replication-planner tests passed
Planner status: hold
Top opportunity: alzheimer + crispr + single-cell
Top replication decision: prioritize
Top peer-review flag: Claim has no supporting evidence artifact.
Weak reproducibility project: circadian-metabolomics-2026
Audit hash: de1b1fa44644dc24...
codec_name=h264
width=1280
height=720
duration=4.000000
git diff --check passed
secret pattern scan returned no matches

AI-assisted with OpenAI Codex; I reviewed and verified the implementation before submitting.