Maintainer Cockpit is a self-hosted dashboard that helps open source maintainers understand and navigate large GitHub pull-request backlogs.
The current supported release is v0.0.1. Maintainer Cockpit is pre-1.0, so review the release notes before every upgrade.
- collects public pull requests from explicit repository collections;
- presents sortable, filterable GitHub facts without an aggregate priority score;
- adds evidence-backed Review Cognitive Load, waiting-on, quality, contribution context, and feature-correlation signals;
- preserves incomplete, unknown, stale, and failed states instead of presenting partial evidence as complete; and
- provides private maintainer workflows for Important PRs, goals, snooze, and ignore.
Maintainer Cockpit supports human judgment. It does not review code, recommend one “best” pull request, write to repositories, or merge and close pull requests. Model endpoints are optional and receive no product credentials, tools, external-URL access, or mutation capability. Personal workflow state is private to the authenticated maintainer.
Requirements are Docker Engine and GitHub credentials that can read the configured repositories.
docker pull ghcr.io/arthursens/maintainer-cockpit:v0.0.1
git clone --branch v0.0.1 --depth 1 \
https://github.com/ArthurSens/maintainer-cockpit.git
cd maintainer-cockpitEdit config/maintainer-cockpit.yaml, replace every example value, and export
the environment variables named by its secret references. The
installation guide provides the complete validation
and container startup commands for the versioned image.
The application should remain bound to localhost. Configure an HTTPS reverse proxy before enabling GitHub login or making the deployment reachable from an untrusted network.
The documentation index routes maintainers, deployment operators, and contributors by task.
- Configuration reference
- Configure GitHub authentication
- Model analysis
- Deployment operations
- Troubleshooting
The docs describe implemented behavior on main. Future work belongs in
GitHub Issues or Discussions until the behavior is available and verifiable.