scrubby_index

Index a repository for Scrubby code review. Scans files and git history; in remote mode, indexes via the GitHub App.

Index a repository for Scrubby code review. The first call against a repo does a full scan: file tree, imports, git history, domain discovery, segment analysis, convention extraction. Subsequent calls can be incremental.

Parameters

Name	Type	Required	Description
`repo_path`	`string`	No	GitHub repo name in `owner/repo` format (e.g. `ScrubbyAI/api`). Required if no `repo_id` is provided.
`incremental`	`boolean`	No	If `true`, only index changes since the last indexed commit. Default: `false`.
`repo_id`	`number`	No	Existing Scrubby repository ID. If omitted, creates a new repository or finds an existing one by name.

Behavior

When called without an existing index, Scrubby:

Scans the file tree, hashing every file and parsing imports.
Sends directory structure, file metadata, and a sample of file contents to the domain classifier.
Builds the domain graph and computes cross-domain connection weights.
Activates relevant global domains (Ruby, React, Testing, etc.) based on dependency files.
Ingests recent git history (no author data, no PII).
Saves a snapshot recording the head SHA, file count, and whether reclassification was triggered.

When called with incremental: true, only files whose content hash has changed since the last indexed SHA are re-processed. Domain reclassification is skipped unless significant changes are detected (e.g. dependency files changed, more than 20 new files added).

Typical usage

In your AI editor:

"Index this repo with Scrubby. It's owner/repo-name."

For an incremental refresh after large changes:

"Re-index incrementally with Scrubby."

When to re-index

Scrubby tracks new commits incrementally on its own once the first index is built. You should explicitly trigger a re-index when:

The repository has been heavily restructured (large directory moves, language additions).
You want to force convention re-extraction after a refactor sprint.
The dashboard shows the index is stale (warning surfaced after very long inactivity).

Errors

Code	Meaning
`repo_not_found`	Either the GitHub App isn’t installed on the repo, or the repo name in `repo_path` is wrong.
`not_authenticated`	OAuth session expired. Reconnect from your editor.
`index_in_progress`	A previous index is still running. Wait a minute and retry.