Features
Repository Review
Run scheduled audits and turn evidence-backed findings into parked work.
Scheduled repository audits
Repository Review runs on the configured cadence. It inspects the repository for maintenance, correctness, security, quality, or architecture findings described by its instructions.
Parked and deduplicated findings
Verified findings become GitHub issues in the configured Project Backlog. They are parked for human prioritization rather than automatically promoted to Ready.
A scheduled audit may validly produce no issue when it finds nothing that meets the configured bar.
Read-only execution profile
Repository Review uses an ephemeral checkout and a repository-review execution profile. It must not modify the checkout. Issue Planner uses its own fixed planning image and does not permit an override.
Frequently asked questions
What triggers Repository Review?
The repository audit becomes due according to the selected schedule.
Where do findings appear?
Verified findings are created as parked GitHub issues in the configured Project Backlog.
Is an audit allowed to find nothing?
Yes. Scheduled-audit contracts permit zero issues when no finding meets the configured standard.
Can Repository Review change the checkout?
No. The run is intended to inspect the repository and fails if it changes the checkout.
Do provider credentials reach the hosted website?
No. Provider API keys and OAuth roots remain in the local runner environment. The hosted control plane receives sanitized availability, model catalog metadata, job events, and outcome evidence.
Can agents in one repository use different models?
Yes. Each built-in or custom agent selects its own provider, catalog model, supported thinking level, and repository tool image.