Skip to main content

sow

Brainstorm goodness-based project ideas with the user — a structured ideation session that starts from service (who is hurting, who do you already serve) rather than opportunity, generates candidate seeds, rough-vets each through the four gates, and plants keepers in the garden. Use when the user wants project ideas, asks "what should I build next", or wants to brainstorm their future good/pure/valuable technical project.

Quellinformationen

Repository
nagisanzenin/seed-protocol
Letzte Quellaktivität
22. Juli 2026 um 03:46
Erkannte Sprache von SKILL.md
Englisch
Sterne
0
Forks
0

Installationsoptionen

Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.

Quelldateien prüfen

Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.

SKILL.md wird angezeigt

SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
sow
description
Brainstorm goodness-based project ideas with the user — a structured ideation session that starts from service (who is hurting, who do you already serve) rather than opportunity, generates candidate seeds, rough-vets each through the four gates, and plants keepers in the garden. Use when the user wants project ideas, asks "what should I build next", or wants to brainstorm their future good/pure/valuable technical project.
argument-hint
optional: a domain, a pain you've seen, or nothing
# /sow — find seeds worth planting Goal: leave this session with 1–3 candidate seeds the user actually feels called to, each rough-vetted, planted in the garden as SEEDLING. Resolve `$ROOT` and `$GARDEN_DIR` as in `skills/seed/SKILL.md` (root waterfall + landmark, garden dir, entry format). Read `$ROOT/doctrine/DOCTRINE.md` and skim `$ROOT/doctrine/EXEMPLARS.md` first. ## Step 1 — Harvest (the raw material is the user's life, not the market) Good seeds are found in what the user has *witnessed and carried*, so harvest before generating. Use one AskUserQuestion round (multiSelect where natural) plus open conversation, covering the five vectors: 1. **The Wound** — pain they've experienced or watched someone experience firsthand. 2. **The Neighbor** — people they already serve and know deeply (family, colleagues, their community, their country's context). 3. **The Lantern** — hard-won knowledge they hold that others stumble without. 4. **The Commons** — a shared resource that should exist and doesn't, that everyone would tend. 5. **The Dignity** — a place where people are currently treated as metrics/cattle (ads, dark patterns, extraction) and deserve a tool that treats them as persons. Also mine what you already know of the user (their projects, skills, past work) — personalize; do not harvest from a blank slate. ## Step 2 — Generate candidates Produce 6–10 candidate seeds across *different* vectors. Each candidate is exactly three lines: - **who**: the served human, specific enough to picture - **gift**: what concretely gets better in their week - **mechanism**: how, in one honest sentence (parties + flows, no adjectives) **Blocked vectors — never generate from these** (except explicitly labeled as a teaching counter-example): The Arbitrage (value from a gap someone else will close), The Trend (built because it's hot, not because someone's served), The Trap (capture-then-monetize), The Skim (interposing in a flow without adding value). ## Step 3 — Rough-vet Run all candidates through the four gates *quickly* — enum verdicts only, one line of evidence each (vocabularies in DOCTRINE.md). Render as a table. Kill obvious REFUSEs out loud, naming which gate failed and why — the kills teach the doctrine. ## Step 4 — Choose AskUserQuestion (multiSelect): which candidates to keep as seedlings — but weight the user's *pull* over the scores: a seed the user feels called to tend beats a seed that merely scores well. Say so. ## Step 5 — Plant Append each keeper to `$GARDEN_DIR/garden.md` in the standard entry format, status `SEEDLING`, gates from the rough-vet. Close by offering next step: deep-vet one now (`skills/vet/SKILL.md`) or stop here.
Auf GitHub ansehen