route

# MasterMind — Wayfinder The front door of substantive work: decide **what to load** before you start, so you pull the 1–2 things a task needs instead of reading the whole brain. It's the skill that operationalizes MasterMind's token economy (`~/.mastermind/engineering/active-field.md`) — and it knows when *not* to run. ## Method 1. **Name the destination.** One line: what is this task actually producing? Scope before routing. 2. **Refuse to over-plan.** If it's a one-line diff, a typo, or obviously covered by what's already in context — say *"no map needed, just do it"* and **exit.** Ceremony on a trivial task is waste (this is the same instinct as `~/.mastermind/engineering/core/rigor.md`'s refuse-list and "keep the always-on layer tiny"). 3. **Detect the field** — `package.json`/configs/file types (usually free and decisive). 4. **Build the load manifest — links, never summaries.** Consult `~/.mastermind/engineering/ROUTER.md` and pick, by matching the task to each node's `route_when`: - which **field-pack file(s)** to load (and their token cost — budget it), - which **skill(s)** fit — pick from the index at `~/.mastermind/skills/README.md` (which also names the isolated-context **agents** in `~/.mastermind/agents/`); never route off a remembered list, - which **core docs** the task genuinely needs (`mindset`/`principles`/`rigor`/`agent-loop`/`product-sense` in `~/.mastermind/engineering/core/`), - which **codebase areas** to read first...