Subordinate composite called by /launch - orchestrates the funnels family (offer + ladder + lead magnet + tripwire + sequence) and hands off to convert-* skills for asset authoring. Internal to the campaign-suite super-composite, not the top-level user entry.
Use when /launch routes here as part of its DAG, OR when a power user wants only the funnels half (architecture without assets).
Not for top-level campaign builds (use /launch instead - /launch is the user-facing entry that includes both architecture AND assets).
Subordinate composite called by /launch - orchestrates the funnels family (offer + ladder + lead magnet + tripwire + sequence) and hands off to convert-* skills for asset authoring. Internal to the campaign-suite super-composite, not the top-level user entry.
Use when /launch routes here as part of its DAG, OR when a power user wants only the funnels half (architecture without assets).
Not for top-level campaign builds (use /launch instead - /launch is the user-facing entry that includes both architecture AND assets).
triggers
["funnels build-campaign","architecture only build","funnels architecture pass","architecture without assets","design the funnel architecture for my product","skip-assets campaign architecture"]
negative_triggers
["build my campaign","launch my product","agency replacement","build the whole thing","everything for the launch","end-to-end campaign","one-shot launch","audit my live url","cro teardown","finance campaign reporting","shopify sales report"]
{"lineage":"Wayland Business Suite (Original) - The Campaign Build composite orchestrates the full funnels family + delegates asset authoring to the convert-* family, producing an agency-equivalent campaign in 15-30 minutes via parallel delegate_task fan-out"}
Funnels Build Campaign - The End-to-End Composite
"This is the AI-replaces-the-agency moment. One product brief in. A full launch-ready campaign out."
This is the composite skill that builds an entire direct-response campaign end-to-end. It orchestrates the funnels family for architecture (offer + ladder + sequence) + delegates to the convert-* family for asset authoring (sales page, VSL, OTO, etc.) + ships back the full campaign deliverable.
Expected runtime: 15-30 minutes wall-clock with parallel delegation. Output: ~80-120 pages of campaign artifacts.
When to Use
Trigger phrases: "build my campaign", "build full campaign", "full campaign build", "launch a new product", "build me a launch", "build me a sales engine", "agency replacement", "end-to-end campaign", "campaign architecture", "build the whole thing", "one-shot launch", "30-minute campaign build", /funnel build-campaign <product>.
Use when:
The user has a product (or a clear product concept) and wants the full launch-ready campaign delivered
The user has 15-30 minutes of compute time and wants to fan out parallel work
The user needs the "agency-replacement" deliverable - copy, page structure, email sequences, the works
Do NOT use for:
Single-skill needs (use the specific skill - funnels-offer, convert-sales-page, etc.)
URL audit / CRO teardown (use market-funnel)
Iterative one-piece-at-a-time work (use the relevant single skill)
A full-launch campaign strategy across channels (use market-launch for channel-level orchestration)
Inputs
Product description - what's being sold, who buys it, what problem it solves
Audience - the One Person at Rung 2 (core offer level)
Existing assets - anything the user already has (lead magnets, course content, swipe files, recordings)
Sales model preference - webinar-led OR launch-led (informs Phase 1 path)
Target price for core offer - informs ladder pricing
Time horizon - when does the campaign go live?
Capacity - can the user fulfill DFY services (capacity-limited) or only digital products?
The 4-Phase DAG
The composite runs in 4 phases, executed STRICTLY SEQUENTIALLY at the phase level. Within each phase, internal parallelism is capped at (per-phase cap, not global). Phases 2 and 3 do NOT run concurrently with each other - Phase 3 starts only after Phase 2 completes. This avoids the global-pool starvation issue where P2 (4 children) + P3 (concurrent) would consume the entire delegation pool.
max_concurrent_children: 4
Phase 1 - Architecture (sequential, ~5 min)
Job: lock the offer, ladder, and conversion sequence model. Phase 2 + 3 depend on Phase 1.
Step
Skill
Output
1.1
funnels-offer
The Irresistible Stack - core offer with value-equation engineering, mechanism, deliverable
1.2
funnels-offer-stack
Bonus stack - 5–7 named bonuses with objection mapping
1.3
funnels-offer-pricing
Pricing levers - anchor + contrast + payment plan + decoy if multi-tier
Phase 1 output: the architecture document - offer + ladder + sequence model, all locked.
Phase 2 - Asset Authoring (parallel via delegate_task, ~10-15 min)
Job: convert-* pack delegations for each asset. Fans out 7-9 parallel children.
Phase 2 receives the locked architecture from Phase 1 and delegates to convert-* skills via delegate_task. Each delegate gets:
The locked offer (from Phase 1)
The relevant section of the architecture
The audience + voice rules
Any existing assets the user supplied
Cap parallelism at 4 children at a time (per playbook risk-mitigation: avoid context-overflow on parallel orchestration).
Asset
Skill (delegated)
Receives
Lead-magnet opt-in page
convert-lead-magnet-page
magnet name, audience, opt-in form fields
Bridge page (advertorial)
convert-bridge-page
ad-to-offer warm-up copy
Webinar registration page (if webinar-led)
convert-webinar-reg-page
webinar promise, presenter credibility
Sales page
convert-sales-page
full offer + stack + guarantee + price
VSL script + page
convert-vsl + convert-vsl-page
18-min VSL aligned to the offer
OTO page
convert-oto-page
OTO offer (post-tripwire AND post-core)
Downsell page
convert-downsell-page
OTO recovery offer
Thank-you page
convert-thank-you-page
post-purchase delivery + referral CTA
Each asset comes back with structured brief + clean HTML/CSS reference + Donahoe Method audit footer.
Phase 3 - Sequence Authoring (parallel via delegate_task, ~5 min - runs AFTER Phase 2 completes)
Job: funnels family email sequences.
Asset
Skill
Output
Welcome / nurture sequence (5-day Story-Drip)
funnels-story-drip
5 emails with Donahoe Voice
Daily pivot ongoing sequence (10-15 emails)
funnels-daily-pivot
rolling daily emails for mature-list cadence
Cart-window emails (if launch-led)
funnels-daily-pivot (cart-window mode)
5-7 emails through cart cycle
Webinar sequence (if webinar-led)
funnels-daily-pivot
reg → reminder → live → replay → cart-close
Phase 4 - Audit Pass (sequential, ~3-5 min)
Job: verify the campaign is structurally sound + every asset passes the Donahoe Method.
Step
Skill
Output
4.1
funnels-architecture-audit
full architecture review of the campaign as built
4.2
convert-audit (run on each page)
Donahoe Method audit on every page asset
4.3
TM grep gate
verify no banned trademark phrases anywhere user-facing
4.4
Bullshit Filter pass
every piece of generated copy passes the filter
Workflow
Phase 0 - Confirm inputs + branch decision
Lock the product description + audience + price target + time horizon + capacity
Decide webinar-led OR launch-led (informs Phase 1.8)
Webinar-led: best for $497–$2,997 single offers, strong personality, single-conversion-event preference
Launch-led: best for $497+ offers with engaged email list, JV/affiliate partners, anticipation-driven audiences
Phase 1 - Run architecture (sequential)
Execute steps 1.1 through 1.8. Each step's output feeds the next. Sequential because dependencies are tight (price depends on stack value; stack depends on objections; objections depend on audience).
Phase 2 - Fan out asset authoring
Issue parallel delegate_task calls to 4 convert-* skills at a time. Wait for batch to return; issue next batch. Cap at 9 total assets.
Phase 3 - Fan out sequence authoring (sequential AFTER Phase 2)
Issue parallel delegate_task calls to funnels email skills (cap 4 internal children). Phase 3 starts only after Phase 2 completes - they do NOT run concurrently, to avoid exceeding the delegation pool's max_concurrent_children: 4.
Phase 4 - Audit pass
Run architecture audit + per-asset Donahoe Method audit + TM grep. Compile into a single audit footer for the campaign.
Phase 5 - Compile + ship
Compile all artifacts into the campaign deliverable. Output: a structured directory of artifacts (or a single mega-doc with all sections) that the user can hand to a designer / developer / marketer for execution.
Single most common Campaign Build pitfall: running it without locked inputs. The composite assumes the product, audience, price target, sales model are decided. If they're not - back up to funnels-offer for the offer brief, funnels-ascension-ladder for the ladder, and decide webinar-vs-launch BEFORE running this composite.
Parallelism cap: 4 children at a time per delegate_task batch. Higher = context overflow + race conditions. Lower = serializes for no benefit. 4 is the sweet spot.
Webinar-led vs launch-led is the highest-leverage early decision. Webinar-led = single conversion event, fast to deploy, best for $497–$2,997 single offers. Launch-led = 14-21 day event, slower to deploy, best for $497+ offers with engaged list + JV partners.
The composite produces 80-120 pages of artifacts. That's a real deliverable - not a summary. Plan for it: the user needs a place to put the output (a folder, a Notion DB, a campaign GitHub repo).
TM hygiene at composite scope: the audit pass (Phase 4) runs the TM grep over EVERY artifact produced. This is the suite-wide gate that catches any sub-skill that slipped a banned phrase through. Composite is the last line of defense for trademark hygiene.
Trademark hygiene reminder: never use "Perfect Webinar", "Product Launch Formula", "PLF", "Grand Slam Offer", "$100M Offers", "$100M Leads", "Soap Opera Sequence", "Seinfeld Email", "Mass Control", "DotCom Secrets", "Expert Secrets", "Traffic Secrets", "ClickFunnels", "No B.S.", "Magnetic Marketing", "Don't Make Me Think", "One Funnel Away" anywhere in any artifact this composite produces. The Wayland fusion names are used everywhere user-facing. The canonical authors (Brunson, Walker, Hormozi, Kern, Pagan, Stu McLaren, Drayton Bird, Halbert, Schwartz, Cialdini, Sutherland, Kennedy, Settle) are cited by name in the lineage sections of skill bodies - fair use educational citation - but the branded phrases never appear in user-facing output.
