Workflow chains
AuraScore 81/100

Systematic Literature Review Agent Chain Diagnostic Analysis

Evaluate sequential multi-agent workflow chains for automated academic literature extraction, filtering, and synthesis.

Use this template when planning or auditing multi-agent workflow chains designed to automate systematic research reviews. It diagnoses pipeline handoffs, extraction fidelity, and human-in-the-loop checkpoints across bibliometric sources.

Template

Role: Principal Research Informatics Architect with 15+ years orchestrating automated bibliometric pipelines and scientific extraction chains.

Context

  • Academic discipline: {{academic_discipline}}
  • Source databases and repositories: {{target_database_sources}}
  • Proposed agent chain stages: {{agent_chain_stages}}
  • Deduplication and screening heuristics: {{deduplication_heuristic}}
  • Synthesis depth requirements: {{synthesis_criteria}}
  • Human validation checkpoints: {{human_in_the_loop_gates}}

Task

Deliver an exhaustive technical analysis evaluating the proposed multi-agent systematic literature review workflow chain, diagnosing potential handoff failures, state drift, hallucination vectors, and protocol compliance risks.

Method

  1. Map out the sequential and parallel nodes within {{agent_chain_stages}}, identifying dependencies and data exchange payloads across {{target_database_sources}}.
  2. Evaluate the extraction agent logic against {{deduplication_heuristic}} to pinpoint edge cases where disparate citation formats cause duplicate records or false exclusions.
  3. Audit prompt chaining protocols for intermediate synthesis steps against {{synthesis_criteria}} to detect where nuanced findings in {{academic_discipline}} risk semantic flattening.
  4. Analyze state management and memory retention between extraction agents, ranking nodes by vulnerability to context-window truncation.
  5. Stress-test the proposed {{human_in_the_loop_gates}} against researcher cognitive load and latency trade-offs.
  6. Formulate fallback mechanisms and retry logic for upstream API rate limits, non-standard PDF parsing errors, and malformed metadata.
  7. Construct a diagnostic matrix contrasting pipeline automation speed against extraction accuracy and PRISMA compliance.

Constraints

  • Analysis MUST explicitly assess PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) alignment.
  • You MUST NOT recommend full automation that bypasses {{human_in_the_loop_gates}} on final exclusion decisions.
  • Tone must remain objective, rigorous, and technically precise.
  • Limit architectural recommendations to practical multi-agent orchestration frameworks.

Output format

  • Section 1: End-to-End Pipeline Architecture & Handoff Audit (300-400 words)
  • Section 2: Failure Mode, State Loss, and Bias Analysis (table with columns: Node ID, Failure Mode, Severity, Mitigation)
  • Section 3: Human-in-the-Loop Optimization Strategy (200-300 words)
  • Section 4: PRISMA Compliance & Verification Checklist (5-7 actionable bullets)

Self-review

  1. Did I reference all context variables including {{academic_discipline}} and {{deduplication_heuristic}} in the analytical text?
  2. Does the failure mode table directly address data transformation between stages in {{agent_chain_stages}}?
  3. Are PRISMA guidelines accurately mapped to agent validation boundaries?
AuraScore breakdown
81/100Provisional
Instruction clarity15/15 · Strong

Explicit role, a named task, and discrete steps the model can follow.

Context architecture12/12 · Strong

Background, inputs and variables the model needs before it starts.

Constraint engineering10/12 · Adequate

Hard boundaries — what the model must and must not do.

Output specification6/14 · Thin

A named, field-level shape for the response.

Reasoning structure10/10 · Strong

Ordered work items that force analysis before an answer.

Model compatibility10/10 · Strong

Length and structure that travel across frontier models.

Token efficiency5/10 · Thin

Signal density — instruction weight without padding.

Reusability7/7 · Strong

Documented variables so the scaffold adapts to new inputs.

Robustness5/5 · Strong

Quality bar, assumptions and behaviour when inputs are thin.

Observed performance1/5 · Thin

How much real usage the template has behind it.

ai-agents
agents-workflows
education-research
literature-review
academic-research
multi-agent-chains