yogsoth-ai/de-anthropocentric-research-engine

What it solves

DARE is an autonomous research orchestration system designed to remove the "human bottleneck" from scientific discovery. It moves beyond simple AI assistants by acting as the primary researcher that autonomously searches literature, identifies research gaps, generates and stress-tests hypotheses, and designs experiments without requiring constant human guidance.

How it works

The system operates as a "pure-skill" architecture consisting of over 900 markdown-based instruction sets (skills) that are executed by Claude Code (CC). It avoids fixed pipelines in favor of an "arsenal" approach, where the AI decides which research packages to invoke and in what order based on the current state of the research.

It uses a strict four-layer command hierarchy to organize operations:

• Campaigns: High-level research phases (e.g., knowledge acquisition, stress-testing) with defined completion and backtrack conditions.
• Strategies: Iteration engines that manage loops (e.g., search-read-reflect) and decide when to stop.
• Tactics: Workflows that combine multiple atomic operations into coherent outputs.
• SOPs: Atomic, single-responsibility operations that interface with external tools via MCP servers (e.g., Semantic Scholar, Brave Search).

Who it’s for

It is built for researchers and scientists who want to automate the discovery process, from initial direction crystallization to the production of executable research specifications.

Highlights

• Autonomous Gap Discovery: Employs 15+ methods to find what a field is missing rather than just searching for known keywords.
• Extensive Ideation Toolkit: Includes 31+ methods such as SCAMPER, TRIZ, and biomimicry for hypothesis generation.
• Adversarial Stress Testing: Uses multi-perspective attacks and "sacred cow hunting" to validate ideas before acceptance.
• Executable Research Specs: Produces machine-readable documents with quantified completion criteria and automatic session recovery.
• MCP Integration: Connects to 7 external servers including Semantic Scholar and AlphaXiv for deep academic data acquisition.