Operations Flow Optimization

npx skills add https://github.com/peter-tu-zynkr/zynkr-skill-builder --skill operations-flow-optimization

Take a messy, human-grown business process and turn it into a clean "ideal flow" — I→P→O-valid on every step, dependency-ordered, and stripped of friction — before anyone builds or automates it. This is the lean-optimization pass that most teams skip: they jump straight from "here's our design" to "let's automate it," and end up automating waste. This skill makes you earn a streamlined flow first. Its canonical slot (course Ch5 §5.5.2, "Redesign the Flow") is right after the as-is→to-be transformation pass has re-chained the steps (operations-transformation Stage 3) and right before drawing on Lucid (product-flow-design) or building — though it also runs standalone on a raw SIPOC / as-is map whenever a flow feels tangled. It is deliberately conservative about re-ordering — it changes sequence only when the change pays for itself.

The one law it enforces: eliminate before you automate. Anchor the spine to I→P→O and dependencies; only then, rarely, re-sequence — and only for less rework, lower risk, or higher throughput.

Step 1 — Collect the flow

Ask the user for (accept whatever they have; don't block on all three):

The process, as steps — a list of the steps in their current order. The redesigned to-be flow from operations-transformation Stage 3 is the ideal input (the canonical Ch5 §5.5.2 slot); a SIPOC map from Stage 1 or a rough bullet list is fine too.
The outcome — what the process must reliably produce when it's done (乾淨的追蹤名單, approved + notified, 月結報表). This is the anchor everything gets tested against.
Known pain — anything that already feels slow, repeated, error-prone, or out of order.