How jairzinhosantos/metaprompting-for-academic-deep-research works
This repository packages a two-phase academic research workflow: a thinking model turns course context and constraints into an optimized research prompt, then a deep-research model uses it to discover and validate papers, formulate problems, and produce structured JSON. The README documents this workflow and its quality criteria; the repository tree provides the prompt files and example result, but no executable orchestration. Relationships shown are limited to those explicitly described in the README.
Main entry points: prompt_for_deep_research.prompt Research Prompt · output.json Research Results · prompt_for_thinking_model.prompt Thinking Prompt
Components 9
Prompt Optimization
- Thinking Model
Research Execution
- Deep Research Model
- Academic Databases
- Paper Validation
- Problem Formulation
Workflow Artifacts
- Thinking Prompt
prompts/prompt_for_thinking_model.prompt - Research Prompt
prompts/prompt_for_deep_research.prompt - Research Results
results/output.json
Other components
- Researcher
Connections 11
- Researcher → Thinking Model: submits context
- Thinking Prompt → Thinking Model: guides analysis
- Thinking Model → Research Prompt: generates prompt
- Researcher → Deep Research Model: submits prompt
- Research Prompt → Deep Research Model: specifies research
- Deep Research Model → Academic Databases: searches papers
- Academic Databases → Deep Research Model: returns candidates
- Deep Research Model → Paper Validation: validates papers
- Deep Research Model → Problem Formulation: formulates problems
- Deep Research Model → Research Results: structures results
- Research Results → Researcher: delivers results