---
prompt_id: c0642d14
version: 1
author: promptadmin
created_at: 2026-07-21T14:04:56.563179+00:00
detected_topics: ['constraints', 'contribution', 'engineering', 'generation', 'limitations', 'literature', 'optimization', 'specialty']
execution_model: openai/gpt-oss-120b
---

<Task>
  Generate a complete research paper (max 5 pages) on "Augmented Prompt Engineering with Retrieval‑Augmented Generation (RAG)" targeting top‑tier venues (ICML, NeurIPS, ICLR).
</Task>
<Constraints>
  <Length>≤5 pages, double‑column, 10pt font</Length>
  <Citations>Only papers from ICML, NeurIPS, ICLR, CVPR, ACL published 2022‑present. Use live internet search; do not fabricate references.</Citations>
  <Style>Academic, third‑person, include abstract, intro, related work, method, experiments, discussion, conclusion, and references.</Style>
  <Focus>Hybrid human‑computer prompt generation, limitations of purely human‑written prompts, model‑specific optimization, and the need for a prompt database.</Focus>
</Constraints>
<Output>
  <Section name="Abstract">[250‑word abstract]</Section>
  <Section name="Introduction">[Motivation, problem statement, contributions]</Section>
  <Section name="Related Work">[Summarize ≥5 top‑tier papers (2022‑2024) on RAG, prompt engineering, hybrid approaches]</Section>
  <Section name="Methodology">[Describe augmented prompt design, retrieval pipeline, hybrid generation, model‑specific tuning]</Section>
  <Section name="Experiments">[Design a factual‑check experiment comparing human, automated, and hybrid prompts; report metrics]</Section>
  <Section name="Discussion">[Limitations, future work, need for a prompt repository]</Section>
  <Section name="Conclusion">[Key takeaways]</Section>
  <References>Generate a BibTeX list matching the citations above.</References>
</Output>
<Retrieval>
  Before writing, perform an internet search for the latest RAG and prompt‑engineering papers (2022‑2024). Summarize each found paper in one sentence and include its citation key.
</Retrieval>