our-claude-skills/custom-skills/96-ourdigital-estimate-engine/SKILL.md

name: ourdigital-estimate-engine description: Method-aware estimate/견적 engine for OurDigital / D.intelligence professional services (SEO, GA4/GTM, education/coaching, digital ads, branding, …). Generates a Korean 견적 (md/xlsx/json) from a generic scope.json using the real company rate card. Use for "OurDigital 견적", "estimate", "quote", "proposal pricing", "cost estimate", "견적서 생성", or when another skill needs to price a service. Costing methods: effort (role×billing×hours), coaching (lesson×hours), procurement (+15%).

OurDigital Estimate Engine

Single source of truth for OurDigital/D.intelligence pricing. Consuming skills map their context into a scope.json and call the engine CLI; the engine owns the rate card + service catalog and renders the 견적.

Costing methods (catalog entry declares its method)

• effort — role_rate × 청구율(0.70) × 표준 업무시간, by module; auto-tier (smb/basic/treatment) by portfolio size + premium-vertical floor; On-page hours scale by sub-brands (cap ×2.0); 제안가 = 합계 절사. Also supports fixed-amount tasks (fixed: Unit Cost — e.g. branding 진단 stages) and monthly retainers (unit: monthly — e.g. ads 대행). Reproduces real SEO (10.5M/25.0M), branding (₩9.0M), ads (₩6.0M/월).
• coaching — Σ(lesson_type_price × hours) over a lesson plan; student discount opt-in. Reproduces real GA4/GTM 코칭 (₩1,570,000). Base prices default; subject×level matrix optional.
• procurement — unit_cost × qty × (1 + 0.15) for non-labor items.

Files

• references/rate_card.yaml — universal: role rates, billing, basis, terms, tiering, scaling, procurement, coaching prices/levels/discounts, tools. Private.
• catalog/<service>.yaml — per-service: service, method, body. All real: seo (effort, role-hours tiers), education (coaching, incl. GA4/GTM course), digital_branding (effort, fixed 진단 stages → ₩9.0M), digital_ads (effort, ₩6.0M/월 retainer; media commission % is per-deal, not in catalog).
• scripts/estimate.py — CLI dispatcher; scripts/methods/{effort,coaching,procurement}.py; scripts/render.py (md/xlsx/json). scope.schema.json — input contract.

Usage

python scripts/estimate.py \
  --rate-card references/rate_card.yaml --catalog-dir catalog \
  --scope scope.json --out-dir <engagement> [--seq N]

scope.json examples (see scope.schema.json):

• effort: {"service":"seo","tier":"auto","signals":{"properties_total":25,"subbrands_total":5,"vertical":"hotel_resort"},"prospect":{"name":"…","audit_date":"YYYY-MM-DD"}}
• coaching: {"service":"education","course":"ga4_gtm_intermediate","students":1,"prospect":{…}}
• procurement: {"service":"seo","method":"procurement","items":[{"label":"SEMrush Guru","unit_cost":330000,"qty":6}]}

Outputs 05_estimate_ko.md, 05_estimate.xlsx, data/estimate.json. The effort estimate.json shape is consumed by ourdigital-presales-seo/build_deck.py.

Consuming from another skill (CLI)

1. Map your context → scope.json (service, tier signals or lessons, prospect).
2. Call estimate.py. 3. Use the rendered 견적; for SEO, feed data/estimate.json to the deck. Example consumer: ourdigital-presales-seo (findings_to_scope.py → engine → 견적 + deck).

Adding / editing services

• New real service: add catalog/<service>.yaml with method + body from a real quote.
• Any new stub should carry _stub: true (the 견적 prints a ⚠STUB banner) until replaced with real quote data. All current catalogs are real.
• Rates change in rate_card.yaml only (single source). Validate against a known real quote.

Output location

Estimates are archived centrally in OurDigital's space: ~/Workspaces/ourdigital-space/estimates/<YYYY-MM-DD>_<prospect|account>_<service>/ (e.g. …/estimates/2026-05-27_shr_seo/). The engine stays --out-dir-agnostic — the caller passes that path. Client engagements keep their audit docs/deck in their own workspace but the 견적 lives here (single quote archive).

Conventions

Korean-first output · 부가세 별도 · 유효기간 14d · 현금 · OD-YYYY-NNN. Don't invent rates — stub and flag instead. Legal entity (주)디인텔리전스 / info@ourdigital.org.