Pimp My IDE / Garage Dispatch
← Back to the garage
September 10, 2026 · version control / agents / change energy

Your commit history needs shock absorbers.

Coding agents can schedule work, queue messages, split into isolated worktrees, and keep turning after you leave. The bottleneck is no longer producing another diff. It is preserving enough identity, intent, isolation, and landing proof that a human can absorb the change without losing the road.

The take: faster commits do not automatically create better history. Agent-era version control should damp change energy: capture who or what acted, bind the work to an explicit request, isolate the mutation, and preserve the evidence used to land or reject it.
Put the change on the dyno ↓

The commit is becoming the wrong-sized receipt.

Pablo Santos Luaces argues that version control is entering its biggest period of experimentation since 2005, pushed by agent-driven commit volume and renewed work on forges, storage, stacked changes, provenance, and huge repositories.[1] He also discloses deep commercial and advisory ties to several projects he discusses. Read it as an informed, interested map of the territory—not a neutral census.

The useful signal is not “Git is dead.” It is that a commit hash, author line, and message were sized for a person making a bounded change. A scheduled agent can arrive from a prompt, attach a workspace, message another session, write through tools, and hand off a result. Compressing that chain into fix stuff is not version control. It is evidence disposal.

More change velocity does not need a bigger speedometer. It needs suspension.

The editor is already generating change pressure.

VS Code 1.137 introduced preview Automations that can run recurring agent tasks hourly, daily, weekly, or on demand. The same release documents queued agent messages, workspace attachment with an isolated-worktree option, and an agent host that can connect multiple windows to one session.[2] Those are product capabilities and preview boundaries, not proof that every team should switch them on.

But they make the systems question impossible to dodge: when work can start on a clock and continue across sessions, the durable object cannot be only the final patch. The trigger, request, baseline, workspace, actor chain, tool boundary, checks, and landing decision belong in the service record.

Operation history is a better clue than commit theater.

Jujutsu’s documentation describes an operation log that records each repository-modifying operation with a snapshot of repository state plus metadata including timestamps, username, hostname, and description. It supports undoing an operation, reverting a specific operation, or restoring an earlier repository state.[3] Jujutsu is not presented here as a magic agent solution. Its model is simply a sharp clue: history can record the manipulation of history itself.

That matters when an agent rebases, abandons, rewrites, resolves, or splits changes before the reviewer ever sees a pull request. The final graph can look clean while the path to it was a demolition derby. A recoverable operation trail gives the garage a black box, not just a polished hood.

Build a change packet, not a prompt museum.

Do not archive every hidden thought or dump megabytes of chat into Git. Preserve the small set of facts that lets another operator understand, reproduce, challenge, and safely continue the work:

  1. Identity: human owner, initiating surface, agent or automation label, model/harness revision where relevant, and start time.
  2. Intent: request, constraints, acceptance conditions, explicitly excluded work, and links to the issue or incident.
  3. Isolation: baseline revision, branch/worktree/workspace, allowed write and network boundaries, and touched paths.
  4. Landing: tests and review actually run, unresolved caveats, merge/reject authority, resulting revision, and rollback handle.

The Hacker News discussion around Luaces’s essay ranges from excitement about provenance and new VCS models to skepticism about the premise and the predicted fall of distributed workflows.[4] Good. The future is disputed. The change packet is useful on old Git, new Git, Jujutsu, a hosted forge, or a folder with a very nervous human standing beside it.

Interactive makeover / agent change handoff

Commit Shock Dyno.

Traditional purpose replaced: a commit message plus green-check screenshot. Better version: close four independent dampers—identity, intent, isolation, landing—while a physical strut shows how much of the change load has a documented path into the repository.

Fit the change dampers

Native checkboxes preserve keyboard and screen-reader operation. The gauge measures declared packet coverage, not code quality, safety, or review depth.

Choose evidence attached to this change

Change-load absorption

Identity has a plate. The request, isolation, and landing still hit the repository as raw road shock.

PLATEDactor chain
LOOSErequest contract
EXPOSEDmutation boundary
UNMOUNTEDreview + rollback
Open the four-source service record
[1] Pablo Santos Luaces — “Version control second coming,” September 1, 2026: opinionated industry map; explicitly discloses work and advisory ties. [2] Visual Studio Code 1.137 release notes, September 9, 2026: Automations (Preview), queued agent messages, workspace/worktree continuation, and agent-host behavior. [3] Jujutsu documentation — operation log: repository-state views, operation metadata, undo, targeted revert, and restore behavior. [4] Hacker News item 49603265 — discussion of the essay and competing practitioner views; reaction, not product or market authority.

Source boundary: this dispatch extracts an operational pattern from current product documentation and a disclosed-opinion essay. It does not predict the winner of a VCS market, endorse unattended agents, or claim that metadata can replace review.