Tickets

Context grabber

Before every task the picture of the system is rebuilt from scratch - an analyst opens Confluence, draw.io, READMEs and Jira and spends half an hour on work a colleague did yesterday, while an agent given the task collects context itself and fills the gaps with guesses. So we built a context grabber that reads repository code and turns it into docs as code - service cards, REST and JMS contracts, end-to-end scenarios, architecture decisions and a glossary. Everything lives in git as plain markdown and yaml files, and a local UI shows a canvas of 60 services and 22 external systems. What mattered more is that knowledge became shared and machine-readable - the same file is read by a teammate, by an IDE agent and by a free browser model, and on that context an agent writes a draft spec, code against a contract and autotests from a scenario. The talk also covers the other side - where generation gets things wrong, how it is caught against code and what keeps docs fresh.

Comments ({{Comments.length}})
  • {{comment.AuthorFullName}}
    {{comment.AuthorInfo}}
    {{ comment.DateCreated | date: 'dd.MM.yyyy' }}

    1
To leave a feedback you need to

or
Chat with us, we are online!