One tab for every repo, container and AI session on my machine
Lorenzo VicinoDev.to (EN Zone)
2 views
My working day spans about ten Git repositories. Here is where their state actually lived:
which trees are dirty → forty terminal tabs
which branches are behind → a Git client, pointed at one repo at a time
which containers are up → docker ps
the AI conversation I had yesterday about the retry policy → somewhere under
~/.claude/projects, in a folder named after a path
So I built repo-control.
npx repo-control ~/projects
It scans that folder for Git repositories and serves a dashboard on 127.0.0.1.
No account, no cloud, no telemetry.
Four things it puts in one tab:
What needs you — one queue, worst first: unhealthy containers, automations that failed,
repositories to pull, commit or push. Click a row, land in that repo already scoped.
Every repo as a workspace — inline diffs, staging, commit, fetch/pull/push, branches with
divergence and safe checkout, a terminal scoped to that folder that survives navigation,
Compose state when there is a Compose file.
Your AI sessions — it reads the histories Claude Code, Codex and Gemini CLI already write
to disk, keeps only the ones belonging to your workspace, and resumes any of them in a real
terminal, in the right folder, with the right CLI. Nothing is copied anywhere.
Automations — drag Git, Docker and terminal nodes into a workflow: fetch everything,
rebuild a stack, run one command across a selection. A dry run prints the exact commands first.
One decision I will not budge on: the browser never sends a path. It sends a repository ID,
the server resolves it against the root it scanned, and refuses anything outside. That is the
whole difference between "a web page that runs shell commands on my machine" and something I
leave open all day.
What it is not: a team tool, and not where you do rebases or conflict resolution — that stays in
your editor. It never calls the GitHub API. Everything it knows comes from git, docker and
the files on disk.
MIT, TypeScript end to end: github.com/LorenzoVicino/repo-control
Now the part I actually want from you. I have built this against exactly one workflow: mine.
Three questions:
Across your repos, what state do you lose track of most?
Is "one queue of what needs you" the right front page, or would you want something else there?
What would make you close the tab in the first minute?
If you already have a system for keeping ten-plus repositories straight, I would rather hear that than a star.
Who Saw the Connection? #04 — Railways × Human Factors × Software
If you spend enough time on a railway platform in Japan, you may notice something that looks oddly theatrical.
A conductor checks the platform, points down it, says something aloud, looks toward another reference point, and repeats
When an application is running, knowing that it is up does not necessarily mean knowing what it is doing.
For example, imagine an API request that takes several seconds to complete. We would want to answer questions such as:
Is the application receiving more requests than usual?
What happened dur
GPT-6 Astra at high reasoning effort produced the implementation I decided to keep. I still plan to use medium effort by default.
High handled the interactions between retries and persisted state more completely, but its implementation run took about 48 minutes against medium's 31. In a separate re