Rail

A flow is a path through skills. Most work follows the main chain.

Main chain: idea → ship

1. /rail-setup — once per repo (all rail docs under docs/monorail/)
2. /rail-align — light grilling → writes align.md
3. /rail-spec — write docs/monorail/<feature>/spec.md from align.md (refuses empty context)
4. /rail-slice — write tasks/NN-*.md with blockers
5. /rail-build — user-triggered as /rail-build <feature-slug>; runs frontier tasks serially in one session by default (implemented inline in the main session; a fresh implementer sub-agent only on a hand-off trigger — unfamiliar/wide territory or degraded context). When the planning-session context is already heavy, open a fresh session for the build. Concurrent builds require separate git worktrees (see rail-build)

Planning chain (auto-continue)

align → spec → slice is one planning chain. When a stage finishes successfully in this session, continue the next stage in the same session by reading and following that skill's SKILL.md — do not stop and ask the user to type the next /rail-*. One exception: spec → slice is gated — after writing spec.md, /rail-spec stops and asks once (continue slicing now / revise the spec / pause) and never slices without that confirmation.

Align writes align.md → continue /rail-spec → gate → continue /rail-slice (on confirm) Standalone /rail-spec writes spec.md → gate; /rail-slice ends the planning chain (suggest /rail-build <feature-slug> only)