Skip to content

Q001-H2 — No Such Complete Prompts Exist

Research R0049 — Landscape Scan
Run 2026-03-31-02
Query Q001
Hypothesis H2

Statement

No complete, usable AI/LLM system prompt implementing a full analytical rigor framework for research has been published. Only narrow-task prompts (abstract screening, checklist verification, single-step extraction) exist in the published literature and prompt repositories.

Status

Partially supported — The evidence strongly shows no complete prompts exist, but partial implementations go beyond simple narrow tasks, making the strict "only narrow-task" claim too strong.

Supporting Evidence

Evidence ID Summary Strength
SRC01-E01 Comprehensive taxonomy found no full-framework prompt Strong
SRC04-E01 Commercial system prompts lack analytical frameworks Strong
SRC03-E01 PRISMA-trAIce is a reporting checklist, not an operational prompt Moderate

Contradicting Evidence

Evidence ID Summary Strength
SRC02-E01 sroberts implemented 3 SATs including ACH — beyond narrow tasks Moderate
SRC05-E01 Agent Laboratory implements multi-phase research pipeline Moderate
SRC06-E01 AI-Researcher implements complete research workflow Moderate

Reasoning

The "no such prompts exist" hypothesis is too absolute. While no complete analytical rigor framework exists as a system prompt, several implementations go beyond narrow tasks: sroberts implements ACH and Key Assumptions Check; Agent Laboratory and AI-Researcher implement multi-phase research pipelines. However, these are implemented in code (Python + LangChain) rather than as deployable system prompts, and they do not implement named analytical rigor frameworks (ICD 203, GRADE, etc.).

Relationship to Other Hypotheses

  • Contradicts H1 (mutually exclusive)
  • Refined by H3 which captures the nuanced middle ground