Feature Request: Persistent Knowledge Base & Instruction Sets

The customer, Vedic, is requesting a new feature in Gistr to enable more efficient and consistent repeated document analysis workflows. The core request is for functionality similar to Gemini Gems or Microsoft Copilot Agents, allowing users to create a reusable AI agent. This agent would include: 1. A permanent, reusable set of instructions (prompt) that can be applied automatically to new tasks. 2. A permanent, reusable knowledge base (reference documents) that the system can refer to for every new analysis without re-uploading. The customer provided an example of evaluating student exam papers (10-100+ papers in a batch) against a fixed set of 5 mathematics textbooks and a detailed analysis prompt. They emphasize that this capability would eliminate the need to re-enter prompts or re-upload reference materials for each new document, significantly improving efficiency and consistency for recurring workflows. This feature is seen as valuable for various roles including teachers, trainers, certification bodies, and researchers. The Gistr admin has acknowledged the request, noting that a 'Skills' feature for reusable instructions is under development, but the integration of a dedicated, persistent knowledge base within a skill presents a unique challenge, which is being logged as a feature request.