Vivaria is METR's tool for running evaluations and conducting agent elicitation research.

Better 'AskUserQuestion' - A lightweight MCP server/OpenCode plugin/Agent Skills + CLI interface which allows parallel AI agents ask questions to you. Be the human in the human-in-the-loop!

[AISTATS2021] Official implementation of "Sample Elicitation"

Running UK AISI's Inspect in the Cloud

TasteRay skills for psychological profiling through natural conversation

MCP as a Judge is a behavioral MCP that strengthens AI coding assistants by requiring explicit LLM evaluations

Security Expert Elicitation of Risks

A modern CLI tool for generating production-ready Model Context Protocol (MCP) servers

A Claude AI skill for asking clarifying questions and resolving ambiguity before starting tasks. Multi-round elicitation, automatic question grouping, and intelligent stopping rules for better first-draft outputs.

Methods for mathematically aggregating expert judgements

elicitr is an R package, used to aggregate elicitation data. The package is in active development and implements functions based on two formal elicitation methods. You provide the data, and elicitr transforms them into easily readable graphics.

Course-ready CLI reference for the Model Context Protocol. Demonstrates all six MCP primitives (tools, resources, prompts, sampling, elicitation, roots) wired to Anthropic's Claude. Adapted from Anthropic Academy for O'Reilly Live Training.

UMLet ISTAR-palette for goal oriented requirements engineering (GORE)

REST bridge to the MCP ecosystem, powered by mcp-use

This is a minimal Model Context Protocol (MCP) server that showcases how to build MCP tools with elicitation capabilities

Typed questions for AI agents — a portable spec, conformance suite, and reference implementation for MCP elicitation.

Shiny app visualiser of four point elicitation data

Production-grade MCP server with OAuth 2.1, elicitation, OTel observability, and self-scanning security controls. Reference implementation for AI agents over financial data.

Plumbline — a portable Claude skill for structured spec elicitation. Turns a vague design intention into a written, signed spec with a verifier for every success criterion, before any solutioning. Strawman-then-interrogate to convergence, a calibrated assumptions ledger, decision-point tripwires, and human sign-off. Advisory mode for decisions.

Enhanced FastMCP server with real-time progress reporting, Pydantic elicitation, context logging & Claude Sonnet 4.5 integration — file operations with BASE_DIR security