Lore Knowledge Base
Lore is a research knowledge repository you have access to via MCP tools. It stores documents, meeting notes, interviews, and decisions with full citations — not just summaries, but the original content linked back to its source. Use it to ground your answers in evidence and to preserve important context from your conversations.
When to Ingest Content into Lore
Push content into Lore using the ingest tool whenever you encounter information worth preserving:
- - After conversations: When a user shares meeting notes, interview transcripts, or important documents, ingest them so they're searchable later.
- External content: When you fetch content from Slack, Notion, GitHub, email, or other systems, ingest the relevant parts into Lore.
- Decisions and context: When important decisions are made or context is shared that future conversations will need.
Always include:
- -
source_url: The original URL (Slack permalink, Notion page URL, GitHub issue URL) for citation linking. - INLINECODE2 : A human-readable label like "Slack #product-team" or "GitHub issue #42".
- INLINECODE3 : The project this content belongs to.
Ingestion is idempotent — calling ingest with the same content twice is safe and cheap (returns immediately with deduplicated: true).
When to Search Lore
Before answering questions about past decisions, user feedback, project history, or anything that might already be documented:
- 1. Use
search for quick lookups. Pick the right mode:
-
hybrid (default): Best for most queries
-
keyword: For exact terms, names, identifiers
-
semantic: For conceptual queries ("user frustrations", "pain points")
- 2. Use
research only when the question requires cross-referencing multiple sources or synthesizing findings. It costs 10x more than search — don't use it for simple lookups.
- 3. Use
get_source with include_content=true when you need the full original text of a specific document.
When to Retain Insights
Use retain (not ingest) for short, discrete pieces of knowledge:
- - Key decisions: "We chose X because Y"
- Synthesized insights: "3/5 users mentioned Z as their top issue"
- Requirements: "Must support SSO for enterprise"
Citation Best Practices
When presenting information from Lore, always cite your sources:
- - Reference the source title and date
- Quote directly when possible
- If a
source_url is available, link to the original
Example Workflows
User asks about past decisions:
- 1. INLINECODE17
- Review results, get full source if needed: INLINECODE18
- Present findings with citations
User shares meeting notes:
- 1. INLINECODE19
- Confirm ingestion to user
User asks a broad research question:
- 1. INLINECODE20
- Present the synthesized findings with citations
知识库
Lore是一个研究知识库，您可以通过MCP工具访问。它存储带有完整引用的文档、会议记录、访谈和决策——不仅是摘要，而是链接回原始来源的原始内容。用它来将您的答案建立在证据基础上，并保存对话中的重要上下文。
何时将内容摄入Lore
当您遇到值得保存的信息时，使用ingest工具将内容推入Lore：
- - 对话后：当用户分享会议记录、访谈记录或重要文档时，摄入它们以便日后可搜索。
- 外部内容：当您从Slack、Notion、GitHub、电子邮件或其他系统获取内容时，将相关部分摄入Lore。
- 决策和上下文：当做出重要决策或分享未来对话需要的上下文时。
始终包含：
- - sourceurl：用于引用链接的原始URL（Slack永久链接、Notion页面URL、GitHub问题URL）。
- sourcename：人类可读的标签，如Slack #product-team或GitHub issue #42。
- project：此内容所属的项目。
摄入是幂等的——用相同内容调用ingest两次是安全且廉价的（立即返回deduplicated: true）。
何时搜索Lore
在回答关于过去决策、用户反馈、项目历史或任何可能已有记录的问题之前：
- 1. 使用search进行快速查找。选择正确的模式：
- hybrid（默认）：适用于大多数查询
- keyword：用于精确术语、名称、标识符
- semantic：用于概念性查询（用户挫折感、痛点）
- 2. 仅当问题需要交叉引用多个来源或综合发现时使用research。它的成本是search的10倍——不要用于简单查找。
- 3. 当需要特定文档的完整原文时，使用带includecontent=true的getsource。
何时保留见解
对于简短、离散的知识片段，使用retain（而不是ingest）：
- - 关键决策：我们选择X是因为Y
- 综合见解：3/5的用户提到Z是他们的首要问题
- 需求：必须支持企业SSO
引用最佳实践
在呈现来自Lore的信息时，始终引用您的来源：
- - 引用来源标题和日期
- 尽可能直接引用
- 如果有source_url，链接到原始来源
示例工作流程
用户询问过去决策：
- 1. search(authentication approach decisions, project: my-app)
- 查看结果，如果需要获取完整来源：getsource(sourceid, include_content: true)
- 带引用呈现发现
用户分享会议记录：
- 1. ingest(content: ..., title: Sprint Planning Jan 15, project: my-app, sourcetype: meeting, sourcename: Google Meet, participants: [Alice, Bob])
- 向用户确认摄入完成
用户提出广泛的研究问题：
- 1. research(task: What do users think about our onboarding flow?, project: my-app)
- 带引用呈现综合发现
