A cockpit for AI coding agents on GitHub issues with auto-triage, multi-step autofix, and live observability.
cezar is worth checking the docs before setup with trust notes worth reviewing. Check agent compatibility and use-case fit before adding it to your workflow.
gh repo view comerito/cezar --webOpen the official repository or website.
Check the README for package manager, auth, and platform requirements.
Try it in a small test task inside your agent workflow.
Cezar is a tool that helps developers manage GitHub issues using AI agents. It automatically triages new issues, runs multi-step workflows to fix bugs, and lets you watch everything happen live. You can pause, cancel, or resume agent runs, and it works with different AI backends like Claude or Codex.
Cezar is an open-source platform that acts as a control center for AI-powered GitHub issue management. It provides a unified model called 'Actions' — each with a system prompt, reference skills, and allowed side-effects — to automate triage, bug fixing, and more. Key features include: auto-triage on every new issue (classifies, labels, sets priority), 15 built-in Actions and skills, multi-step autofix workflows (verify, root-cause, fix, review, open PR), live observability via a cockpit UI with pause/cancel/resume controls, effect-scoped GitHub access (only allowed mutations like label, comment, close), and support for multiple AI backends (Anthropic API, Claude Code CLI, Codex CLI). It can run in the cloud or on self-hosted runners, and also includes a solo-use CLI for local JSON stores. Built with TypeScript, it uses a monorepo structure with packages for core, runner, and cockpit.
Looks usable, but maintenance, license, or security notes deserve a closer look.
Last commit was about 14 days ago.
12 GitHub stars indicate community interest.
2 open issues signal maintenance load.
MIT license detected.
Automatically triage incoming GitHub issues by classifying, labeling, and setting priority.
Run multi-step autofix workflows that end in a draft pull request.
Monitor and control multiple AI agent runs in real-time via the cockpit UI.
Detect duplicate issues before they clutter the backlog.
Use custom skills from the repository's .ai/skills/ directory to tailor agent behavior.
Requires GitHub API token with appropriate permissions; ensure token is stored securely.
AI agents can modify issues (label, comment, close, assign); review allowed effects carefully.
Self-hosted runners may expose API keys if not properly configured.
12
Stars
1
Forks
2
Issues
MIT
License
A CLI tool to create and hide GitHub comments using the GitHub REST API.
AI-powered GitHub automation for issue triage, PR review, labeling, and security scanning.
Local GitHub Actions runner for AI agents — test CI/CD workflows locally before pushing.
3 security/trust notes recorded.
Setup difficulty is 3/5.