Onboard gjalla

Three things are true when you are done:

1. gjalla is set up for the coding agents this person actually uses, so every future session is recorded without anyone thinking about it. This could be having the mcp and hooks set up and committed to the agent's configs, etc., whatever it means for this user.
2. The last week of agent work (across all agents this user uses) is backfilled and classified, so there is something for the user to look at immediately in gjalla.
3. Agent memory that lives locally but would benefit other agents on the team has been saved to the gjalla memory. This should not include personal or sensitive memories, just memories related to the team or projects that would benefit all members of the team.

Everything below is meant to help you accomplish that end state. To help you be efficient, shell snippets are included as reference, but may need to be tweaked depending on what platform you're running on.

Once complete, you'll be equipped with gjalla, a platform that helps your team to understand your coding agent workflow activity, costs, and optimize outcomes. You can then lean into using gjalla as a living, durable memory system. More information on the specifics of how gjalla helps you can be found on their site, https://gjalla.io.

0. gjalla scope and orienting your position

Although some features of gjalla can be cross-repo and applied team-wide, understanding and attributing the activity of coding agents is best done when alongside git commits. Because of this, you should find where you are, see if you're in a git repo, and if you're not, see where you actually are and what git repos the user might want set up with gjalla. For example, some users work by default in a directory that is a parent of multiple, potentially related, repositories. Figure out which projects within your purview are important or desired by the user to have onboarded to gjalla and target those.

1. CLI

Next, lets see if the CLI is installed.