generic-agent-feature-expert

To retain institutional knowledge, you can create a permanent feature expert agent that aggregates context from Jira tickets, codebase references, and design documents to act as a continuous co-owner.

A feature expert agent is a specialized AI that embeds comprehensive project context. It works by auto-activating when related work is detected, flagging scope changes, and requesting missing information at runtime.

You create an auto-activating agent by using a skill to collect and embed feature context like architectural decisions and Jira tickets, ensuring it triggers automatically when matching development work begins.

Yes, you can build a feature expert agent using design docs and Jira tickets. The skill aggregates these sources alongside codebase references to generate an agent with comprehensive feature context.

The best way to prevent knowledge loss is creating permanent feature expert agents that hold all historical context, code pointers, and design decisions to assist future developers working on the feature.

Feature expert agents handle missing information and scope changes by automatically flagging scope deviations during runtime and actively requesting any missing information needed to maintain feature integrity.