| name | seo-growth |
| version | 1.0.0 |
| description | SEO and conversion growth router: technical SEO, content SEO, landing pages, funnels, CRO, trust patterns. Topics: conversion-optimization, funnel-design, indexability-validator, landing-page-builder, landing-pages, seo-architecture, seo-content, social-proof. |
| params | {"topic":{"enum":["conversion-optimization","funnel-design","indexability-validator","landing-page-builder","landing-pages","seo-architecture","seo-content","social-proof"],"required":true,"infer":"from user request; ask only if ambiguous"},"depth":{"enum":["quick","deep"],"default":"quick"}} |
| routes | {"conversion-optimization":"references/conversion-optimization.md","funnel-design":"references/funnel-design.md","indexability-validator":"references/indexability-validator.md","landing-page-builder":"references/landing-page-builder.md","landing-pages":"references/landing-pages.md","seo-architecture":"references/seo-architecture.md","seo-content":"references/seo-content.md","social-proof":"references/social-proof.md"} |
seo-growth
Router skill. Resolve topic, then Read EXACTLY ONE playbook from routes:.
Never load the whole cluster — one topic, one file. [DOC]
Inputs / Outputs
- In:
topic (one enum), depth (quick|deep), the growth goal + any
URL/page/funnel context. Route to the narrowest topic the request names. [INFERENCE]
- Out: the single playbook's deliverable.
quick → essentials; deep →
apply exhaustively with verification at each step. [DOC]
Topic disambiguation [INFERENCE]
- landing-page-builder = build a new page end-to-end; landing-pages =
patterns/critique for existing pages.
- seo-architecture = structure/indexing/internal links; seo-content =
on-page copy/keywords; indexability-validator = audit crawl/index status.
- conversion-optimization = CRO experiments on a page; funnel-design =
multi-step journey; social-proof = trust elements.
Spine
Discover → Analyze → Execute → Validate. Don't skip Validate even on quick. [DOC]
Quality gates [DOC]
Constitution v6.0.0 enforcement; evidence tags (Alfa core set, single family);
script-first rule (prefer a script over manual steps).
Routing + acceptance rubric and signal→topic matrix live in assets/ (see assets/README.md).
Acceptance (route is correct when)
- Exactly one playbook Read; topic matches an enum verbatim. [DOC]
- Deliverable matches the playbook's contract; gates satisfied. [DOC]
Anti-patterns / self-correction [INFERENCE]
- Reading multiple playbooks "to be safe" — pick one or ask once.
- Inventing metrics/traffic figures: tag derived numbers, never fabricate.
- Request spans 2+ topics or topic is ambiguous → ask, don't guess. Missing
playbook for the resolved topic → stop and report.
