pipelineslisted

# pipelines Reads `~/.claude/masterof/gates/pipelines.txt` (the pre-rendered index for this gate — NOT registry.json, which is ~16k tokens to Read). Uses the same lookup mechanics as the Activation Protocol in `../../SKILL.md` (read registry, try to match the request first), with one structural difference: ## Single-select, not multi-select Every entry here is a complete, self-contained workflow (a full design pipeline, a full GSD phase-execution mode, an image-to-code conversion) — the kind of thing where running two at once doesn't compose, it conflicts. So: 1. Read the `"pipelines"` array. 2. Try to match the user's request to exactly one entry first — if confident, say so directly ("이 작업엔 `<name>` 파이프라인을 쓸게요") and `Read` its `path`. No list. 3. If more than one pipeline plausibly fits (e.g. both `gsd-quick` and `gsd-mvp-phase` could handle a small feature), or the user asks what's available: print the list and ask the user to pick **one**. Never activate two pipelines for the same request — if the task genuinely needs more than one, that's a sign it should be broken into separate requests, not run as two pipelines at once. 4. `Read` only the chosen entry's `path`, confirm, then proceed. Scope: one-shot, in-context only, same as every other gate. ## Adding to the roster `check-skills` routes new pipeline-shaped skills here — see its Step 1 for the "does this run one complete pipeline end-to-end, or does it support a task" test used to decide `pi