How settings are resolved
For any setting, the first source that provides a value wins:- An environment variable, if set.
- The stored value in
~/.atatus/harness/config.json. - The built-in default.
The configuration file
The installer writes~/.atatus/harness/config.json. A typical file looks like this:
harnesses, so two tools on one machine can report to different
projects. Values outside harnesses, such as user_id, apply to every tool.
Core environment variables
These are the settings you are most likely to use:Capture settings
These three control what content is recorded. Privacy explains them in full:Troubleshooting settings
Use these when something is not working:ATATUS_TRANSCRIPT_WAIT_MS exists because some tools write their transcript file just after the hook
fires. Without a short wait, the last response is occasionally missing its token counts. Raise it if
you see turns with no token data on a slow machine.Project naming
The project name is the grouping key, so it decides how your data is organised. Pick a convention before rolling tracing out to a team.Changing the project name creates a new project rather than renaming the old one, and past traces
stay with the old name. Decide the convention first, because splitting or merging history later is
not possible.
ATATUS_USER_ID instead of
changing the project name. That keeps one project while still letting you filter by person.
Set variables per tool
Environment variables apply to whatever process reads them, so where you set them decides their scope. To trace one session differently, export the variable in that shell before starting the tool:~/.atatus/harness/config.json or re-run the installer.