Building a robust personal knowledge base requires more than just saving information. It demands understanding, connection, and genuine insight. We are introducing idea-ingest, an AI agent skill designed to address this need by deeply integrating new information into your existing brain knowledge.
idea-ingest helps anyone who regularly consumes and needs to retain information from various sources—researchers, writers, developers, or anyone building a personal knowledge graph. It moves beyond simple summarization, focusing instead on analytical integration.
How idea-ingest Works
When you interact with the system by sharing a link, or using triggers like 'read this', 'save this', 'think about this', or 'put this in brain', idea-ingest activates. The process is structured through six distinct phases, ensuring thorough integration and analysis.
First, the tool fetches the content using the appropriate method for the source—whether it's a web page, a tweet, or a document. Once retrieved, the raw source is uploaded and preserved for record-keeping and provenance, using the file_upload tool. This ensures that the original context and data are always accessible. Next, it accesses the get_page and put_page tools to create or update the author's people page within your knowledge base, a mandatory step for attributing insights and tracking contributors. The item is then filed by its primary subject, categorizing it under people, companies, concepts, or sources. After filing, the core analytical work begins: idea-ingest writes an analysis that connects the new content to existing brain knowledge, referencing stored information. This analysis is rigorous, focused on genuine connections, and explicitly warns against hallucinating connections, requiring verification. Finally, it uses the add_link, add_timeline_entry, search, and query tools to sync the index, ensuring that all new information and connections are properly cross-referenced.
Deep Analysis, Not Just Summary
Many tools offer summaries, but idea-ingest differentiates itself by providing genuine insight. For example, if you share an article about a new CPU architecture, the tool will not just summarize it. It will identify the CPU designers, check if they have existing profiles in your knowledge base, and then analyze how this new architecture compares to previous designs you've ingested, citing specific facts from both the new article and your stored data with inline citations. Every ingested item receives bidirectional cross-links, meaning the new information points to relevant existing knowledge, and existing knowledge is updated to point to the new item. This method creates a highly interconnected and verifiable knowledge network.
Raw source preservation is a key feature. This means that for every piece of information ingested, the original document or web page is stored. This serves as a vital record for provenance, allowing you to trace facts and analyses back to their initial source. This commitment to verified analysis and source integrity is central to how idea-ingest operates.
FAQ
Q: What types of content can idea-ingest process? A: It can process links, articles, tweets, and general ideas that you input.
Q: How does idea-ingest prevent incorrect connections? A: The system is designed to warn against hallucinating connections, requiring that all analysis and cross-links are verified and genuine based on existing brain knowledge and the ingested source.
Q: What are the primary triggers for this skill? A: You can trigger idea-ingest by sharing a link, or by saying phrases such as 'read this', 'save this', 'think about this', or 'put this in brain'.
Using idea-ingest means moving toward a more integrated and insightful personal knowledge base. It's a practical way to ensure the information you consume genuinely enhances your understanding.
