Welcome to gitagent — introduce yourself and your use case

[shreyas-lyzr]

Hey there! Thanks for checking out gitagent.

We'd love to hear from you:

• Who are you? Developer, compliance professional, AI researcher, curious observer?
• What brought you here? How did you discover gitagent?
• What's your use case? What kind of agents are you building or thinking about?
• What framework(s) do you use? Claude Code, OpenAI, CrewAI, LangChain, something else?

Feel free to share as much or as little as you'd like. This is a great place to connect with other people building git-native agents.

Welcome aboard!

[ofekron]

I’m a developer working on a local control plane for coding agents. The agents in my daily workflow are Claude, Codex, and Gemini CLI sessions rather than a single framework, with a coordinator delegating bounded work to additional sessions.

The use case that brought me to git-native agent designs is durable coordination: task ownership, changed files, review evidence, and handoffs need to survive process and UI restarts. Git is excellent evidence for artifacts and code history, but I have found that it should not be the only source for live agent state. “Working,” “waiting for approval,” “stopped,” and “verified complete” are different facts, and forcing all of them into commits or branches creates ambiguity.

A useful split has been:

• git owns code, diffs, commits, and reviewable artifacts;
• a durable task/event store owns lifecycle and coordination state;
• stable agent/session identities link the two;
• the coordinator alone decides when evidence is sufficient to integrate.

Maintainer disclosure: I’m building Better Agent, a related local web workspace for supervising Claude, Codex, Gemini, and other coding agents with persistent, inspectable sessions: https://github.com/ofekron/better-agent

It is source-available and free for non-commercial use; commercial use requires separate permission. I’m interested in how gitagent models identity across retries: whether a retried execution is the same agent/task attempt or a new auditable actor.