@algora-pbc /claim #16

Summary

Implements a self-contained Research Assistant Evidence Suite for the AI-Powered Research Assistant Suite bounty.

This adds deterministic, locally runnable assistant workflows for:

• auto peer-review reports with domain-aware scoring and evidence-linked findings
• reproducibility checks for artifacts, pinned environments, run commands, expected outputs, and prior attempts
• research-gap discovery ranked by novelty, unresolved questions, replication scarcity, and researcher fit
• a combined dashboard payload and lightweight local API/demo server

The implementation intentionally avoids external API keys, hosted model calls, and dependency installs so maintainers can review it quickly.

Requirement Mapping

• Auto peer review reports: buildPeerReviewReport
• Clarity/coherence checks: section and abstract checks
• Statistical/methodological red flags: sample-size, uncertainty, and randomization checks
• Missing citations/scope misalignment: citation coverage and claims-vs-evidence checks
• Adaptive domain templates: molecular biology, clinical trials, and computational research templates
• Reproducibility checker: artifact, environment, command, expected-output, seed, and prior-attempt reporting
• Research gap finder: findResearchGaps
• Personalized opportunity feed: profile/capability matching
• Suite dashboard: buildAssistantDashboard

Full mapping is in research-assistant-evidence-suite/docs/requirement-map.md.

Verification

cd research-assistant-evidence-suite
npm run check
npm test
PORT=4317 npm start

API smoke tested locally:

• /api/dashboard
• /api/review
• /api/reproducibility
• /api/gaps

Notes

Prepared with OpenAI Codex assistance. The implementation and tests were reviewed and run locally before submission.