Routa

The deciding trait is that the work has a shape other people can read. Goals, the tasks under them and the review state of each one sit on a board rather than in your scrollback, so a colleague can pick up something you started without an hour of explanation. That is the part every other tool in this category leaves to a chat message.

What it costs is process. There is a board to keep current, and boards go stale when only one person is looking. Pick it when the work is shared. Pick a terminal harness when it is not.

A shared record is only as good as its worst contributor. Supporting several coordination protocols widens the set of agents that can attach and guarantees they attach unevenly, because the protocols differ in what they can express about progress, evidence and failure. The documentation names all three and describes no common minimum, so a board can look complete while missing whatever one connection could not say.

What it does right is put worktrees underneath. Parallel work has real isolation rather than a convention.

1. Treating evidence as a first-class object rather than a paragraph in a summary is the decision that distinguishes an auditable system from a narrated one, because the artefact supporting a claim is stored beside the claim. 2. Carrying review state in the same model means approval is a recorded transition rather than a message somebody remembers sending.

2. Nothing enforces that an agent produce evidence before a task advances, so the completeness of the chain rests on convention. The model is right and the guarantee is absent.

An audience that follows a person is a genuine asset and an unusually fragile one. Nearly two thousand stars arrived because of who published this, which is cheaper than any marketing budget and worth nothing to anybody else, since the same audience would follow the next repository rather than this one.

Moat: the author's standing, which is not an asset a company can hold. Likely path: continued publication at the author's pace, with the delivery-process ideas absorbed by tools that have sales teams. Position: use the board, and keep the worktrees reproducible without it.

A single service my platform group runs is a completely different proposition from a desktop application on every machine. One deployment, one upgrade, one place to back up, and the engineers reach it through a browser rather than through an installer. That is the shape my organisation already knows how to operate.

What is missing is identity. There is no single sign-on, no provisioning and no audit trail, so the board records what happened without recording who authorised it. Approved with conditions: it sits behind our authenticating proxy, and the review column is not treated as an approval record until it does.

Shipping the command-line tool to two package registries is a small thing that says the authors expect people to install it in different ways, and a permissive licence means a fork stays legal whatever happens to the project. The protocol support is the part I care about: my existing servers attach as themselves rather than through a shim I would have to maintain.

The stack is a compiled backend under a JavaScript front end, so both halves are readable and neither is a black box. Inference is elsewhere, as usual.