One command
/speckit.axi.review (/speckit-axi-review in skills-based
agents such as Claude Code) opens the review surface for the active feature, collects
your annotations and comments, and applies the queued feedback to the markdown. It does
not generate new files; it edits the feature's existing documents in place.
| Command |
What it does |
/speckit.axi.review |
Open a feature's markdown in a browser review surface, collect annotations and
comments, and apply the queued feedback to the canonical .md files.
Optionally name the feature directory, for example
/speckit.axi.review specs/001-my-feature; otherwise the agent uses
the feature Spec Kit marks as active, then your branch, then the most recent
specs/* directory. Skills-based agents name it
/speckit-axi-review (Claude Code) or
$speckit-axi-review (Codex).
|
Spec Kit also offers a review through two optional hooks. After
/speckit.specify, the after_specify hook offers to open
spec.md. After /speckit.plan, the after_plan hook
offers to open plan.md and the other plan artifacts. Accept the offer, or
run the command later.
Under the hood, /speckit.axi.review drives a small zero-dependency Node
tool, .specify/extensions/axi/templates/web-review/axi-server.mjs, that
binds only to 127.0.0.1. The agent uses these subcommands to bridge your
browser and the markdown:
| Subcommand |
Purpose |
start <feature-dir> |
Open or reuse a session and launch the browser. |
poll <feature-dir> |
Wait for sent notes, then print them as TOON. |
reply <feature-dir> "text" |
Show a message to the reviewer. |
end <feature-dir> |
End the session. |
stop [feature-dir] |
Shut the server down. |
Run any subcommand with --help for its flags and examples. You never call
these by hand; the agent does, while you stay in the browser.