Bedrock Knowledge Base Retrieval Review

ПродвинутыйdataМинимальный контекст: 64K
Открыть источник

Reviews and plans citation-aware retrieval over Amazon Bedrock Knowledge Bases. Produces a least-privilege access plan, knowledge-base discovery criteria, data-source filters, reranking decision, evaluation set, and an evidence-preserving response contract without inventing source content.

Варианты использования

  • Reviewing a Bedrock Knowledge Base retrieval architecture
  • Defining citation and evidence requirements for RAG answers
  • Choosing data-source filters and reranking behavior
  • Preparing a retrieval quality evaluation set

Пример промпта

Review this Amazon Bedrock Knowledge Base retrieval use case.

Inputs:
- Knowledge Bases and data sources: [describe]
- User questions: [examples]
- Required citations: [format and precision]
- AWS region and access boundary: [details]

Produce:
1. A read-only discovery and query plan.
2. Least-privilege IAM actions and resource scope.
3. Data-source inclusion/exclusion rules.
4. A reranking decision with regional and cost constraints.
5. An evaluation set with retrieval, citation, and refusal checks.
6. Failure modes. Never fill missing evidence with model knowledge.

Рекомендуемые модели

Совместимые инструменты

claude-codecursorkirogemini-cliany

Модальности

Вход: text, file
→
Выход: text

Похожие Skills

Автор

OpenModels Community

@openmodelsrun

Издатель

Проверенный источник

OpenModels Community

Адаптировано по указанному источнику

Первоисточник ↗