What problem does it solve? Lab administrators receive many join-the-lab Google Form responses and must manually read, rank, and decide which applicants need human review. This Skill automates the observational classification of those responses so admins get consistent labels, confidence scores, and review questions without reading every submission. ## Core Features & Use Cases - Response Classification: Labels applicants as strong-match, needs-review, missing-information, or not-current-fit using the adminbot_classify_join_form_response tool. - Evidence and Review Questions: Returns confidence scores, evidence pointers, and targeted review questions for each applicant. - Safety Guardrails: Flags missing availability, unclear research interests, and duplicate submissions while avoiding inference of sensitive or protected attributes. - Use Case: An admin receives 40 lab application responses and asks the assistant to triage them; each response gets a label, confidence, and review questions, and borderline cases are handed to a human admin for the final decision. ## Quick Start Classify the latest join-the-lab Google Form responses and list which applicants need human review.
