Commands

fab-kit exposes two surfaces:

1. Slash commands for AI agents (Claude Code, Codex) — /fab-new, /fab-continue, etc.
2. CLI commands for humans and scripts — fab change, fab status, etc.

Slash commands wrap a pipeline of CLI calls; the CLI is what runs underneath.

Slash commands (for agents)

Command	Purpose
/fab-new	Start a new change — creates intake, activates it
/fab-continue	Advance to the next stage (spec, apply, review, hydrate)
/fab-ff	Fast-forward through hydrate, confidence-gated
/fab-fff	Full pipeline including ship + review-PR
/fab-clarify	Refine the current stage’s artifact
/fab-status	Show current change state
/fab-switch	Change which change is active
/git-pr	Commit, push, and open a PR
/git-pr-review	Process PR review comments

Each command’s behavior is defined in .claude/skills/<command>/SKILL.md — plain markdown the agent reads.

CLI commands (for humans)

fab change

fab change new --slug auth-middleware-jwt
fab change list
fab change switch <id-or-name>
fab change rename <new-slug>
fab change archive <id-or-name>

fab status

fab status                              # current change, all stages
fab status finish <change> <stage>      # mark stage done, activate next
fab status reset <change> <stage>       # revert a stage
fab status set-change-type <type>       # fix | feat | refactor | docs | …

fab score

fab score <change>                      # compute SRAD confidence
fab score --check-gate <change>         # exit non-zero below threshold

fab preflight

fab preflight                           # validate init + resolve active change
fab preflight <change>                  # override active change (transient)

fab log

fab log command <skill> <change>        # best-effort telemetry

---

For the full CLI reference (every subcommand, every flag), see the GitHub README.