skills.rest
RorySullivan1
RorySullivan1COMMUNITY
Explore all skills by this author →
GitHub

knowledge-agent-authoring

Design knowledge agents that retrieve source-backed answers from a defined corpus.

Updated May 12, 2026
Search Tags:#curation#schema#retrieval#provenance#grounding#knowledge-agent#life-cycle
One-click install
npx skills add https://github.com/RorySullivan1/claudeBrain --skill knowledge-agent-authoring
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: knowledge-agent-authoring
Source: https://github.com/RorySullivan1/claudeBrain/tree/main/.claude/skills/knowledge-agent-authoring
Command: npx skills add https://github.com/RorySullivan1/claudeBrain --skill knowledge-agent-authoring

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill guides teams to design knowledge agents that curate corpora, maintain context, and retrieve answers with provenance, reducing hallucinations and ensuring trust.

Core Features & Use Cases

  • Grounding is non-negotiable: agents answer from the corpus with provenance.
  • Return the answer, not the corpus: concise, cited results.
  • Retrieval with multi-strategy: keyword, semantic, and structure-based search with ranking.
  • Living knowledge base: curation, freshness, deduplication, provenance.
  • One corpus, one schema: align agent to a single corpus with known index.
  • Use cases: building librarian/retriever agents for docs, memory, or knowledge bases.

Quick Start

Outline a plan to design a knowledge agent that curates and retrieves from a single corpus with grounding and provenance.

Frequently Asked Questions about knowledge-agent-authoring

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design a knowledge agent that retrieves answers with provenance?▼

To design a knowledge agent with provenance, mandate grounding rules that force the agent to answer strictly from a defined corpus using multi-strategy retrieval. This ensures concise, source-backed results without fabricating details.

What is the best way to prevent hallucinations in a retrieval-first agent?▼

Preventing hallucinations in a retrieval-first agent requires non-negotiable grounding, meaning the agent must only return cited answers from a curated corpus. Applying multi-strategy retrieval ranking also enforces this reliability.

How do I manage the life-cycle of a knowledge base for an AI agent?▼

Managing a knowledge base life-cycle involves continuous curation, freshness checks, and deduplication. Aligning the agent to a single corpus with one known schema ensures the retrieved information remains reliable and maintainable.

Can I use multiple retrieval strategies for a single document corpus?▼

Yes, you can apply keyword, semantic, and structure-based search strategies together. Using a multi-strategy retrieval approach with ranking over a single schema ensures accurate and source-backed answers from your document index.

Does a grounded knowledge agent return the full source documents to the user?▼

No, a grounded knowledge agent returns the answer, not the corpus. It provides concise, cited results with provenance while keeping the underlying indexed documents, memory, or stored information hidden from the user.

Related Skills

Explore More Software Engineering Skills→
diegosouzapwdiegosouzapw
108

rag-engineer

Design retrieval-augmented generation pipelines with measurable retrieval quality and grounded citations.

Community
Advanced
arbgjrarbgjr

rag-curator

Index and validate RAG corpus knowledge nodes using standard-library Python scripts.

Community
Intermediate
ricardorochericardoroche
2

rag-design-patterns

Standardizes chunking, vector database integration, embedding management, and retrieval reranking for RAG pipelines.

Community
Intermediate
sergeyitalysergeyitaly
1

prompt-context-hooks

Configure hook-based context grounding strictness for Claude Code sessions.

Community
Intermediate