boost-skill-cli 1.2.17__tar.gz → 1.2.19__tar.gz

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/.vale/styles/config/vocabularies/boost/accept.txt

@@ -1,8 +1,10 @@
+ABM
 agentic
 Algolia
 allowlisted
 argparse
 Arize
+Attio
 Bagisto
 bmad
 bookended
@@ -11,8 +13,10 @@ CI's
 CLIs
 Codecov
 codespell
+ColdIQ
 config
 configs
+CRM
 cyclomatic
 Dependabot
 deps
@@ -29,22 +33,26 @@ eval
 evals
 freemium
 frontmatter
+GEO
 GIF
 gitignored
 GitOps
 glassmorphic
 Grafana
+GTM
 gzipped
 hardcode
 Helm
 hotspots
 Huashu
+HubSpot
 JetBrains
 jonnyeclectic
 jonnyeclectic's
 Kotlin
 langchain
 Langfuse
+LeadMagic
 lookups
 lowercasing
 Magento
@@ -75,12 +83,15 @@ reranking
 rescan
 rescanned
 rescans
+RevOps
 ruleset
 Rulesets
 rulesets
+Salesforce
 sandboxed
 sandboxing
 scrollbar
+SEO
 Shopify
 SonarCloud
 stderr
@@ -120,5 +131,6 @@ WooCommerce
 wordmark
 worktree
 worktrees
+Zapier
 zizmor
 zsh

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: boost-skill-cli
-Version: 1.2.17
+Version: 1.2.19
 Summary: boost — Homebrew for AI coding skills
 Author: Jonathan Reyes
 License: GNU GENERAL PUBLIC LICENSE
@@ -818,15 +818,15 @@ rather than failing: no extra, no key, or no built store each fall back to BM25,
 and `boost doctor` reports which engine is actually serving and why.
 
 Want the whole ecosystem instead of the starter set? boost ships a **curated
-registry catalog** — 460+ classified GitHub registries of skills, Cursor/Windsurf
+registry catalog** — 480+ classified GitHub registries of skills, Cursor/Windsurf
 **rules**, and Claude Code **workflows** (slash commands & subagents),
-collectively indexing ~28,000 scannable items. Every `est_items` is *measured*
+collectively indexing ~30,000 scannable items. Every `est_items` is *measured*
 from the repo's file tree, not estimated, so `--limit` ranks by real size — and
 counted once per item however many agents a repo renders it for, so a pack that
 ships a copy in `.claude/`, `.cursor/` and twelve other dotdirs is not credited
 with fourteen skills (`scripts/measure_registry.py`).
 
-Six domains are curated end to end, each tappable on its own:
+Seven domains are curated end to end, each tappable on its own:
 **`ai`** (RAG, evals, tracing, model serving — official OpenAI, Anthropic,
 LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 **`architecture`** (DDD, clean architecture, system design, C4/ADR),
@@ -834,8 +834,13 @@ LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 TUI/dotfiles — including Impeccable, taste-skill and Huashu Design),
 **`java`** (Spring Boot, Kotlin/JVM, Quarkus, JetBrains, Camunda, Vaadin),
 **`ecommerce`** (Shopify, Magento, WooCommerce, Spree, Bagisto, PrestaShop,
-Stripe, Algolia), and **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
-networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs):
+Stripe, Algolia), **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
+networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs),
+and **`marketing`** (marketing, CRM, email campaigns and customer outreach —
+20 registries, ~2,400 measured items: cold-email and ABM sequences from
+ColdIQ, CRM setup and hygiene for Salesforce/HubSpot/Attio from LeadMagic,
+B2B RevOps lead routing, Zapier's GTM cheat codes, plus paid media, SEO/GEO
+and lifecycle email):
 
 Smaller categories fill in around them. **`efficiency`** collects the packs whose
 items exist to make an agent emit *less* — less code (`ponytail`) or fewer output

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/README.md

@@ -100,15 +100,15 @@ rather than failing: no extra, no key, or no built store each fall back to BM25,
 and `boost doctor` reports which engine is actually serving and why.
 
 Want the whole ecosystem instead of the starter set? boost ships a **curated
-registry catalog** — 460+ classified GitHub registries of skills, Cursor/Windsurf
+registry catalog** — 480+ classified GitHub registries of skills, Cursor/Windsurf
 **rules**, and Claude Code **workflows** (slash commands & subagents),
-collectively indexing ~28,000 scannable items. Every `est_items` is *measured*
+collectively indexing ~30,000 scannable items. Every `est_items` is *measured*
 from the repo's file tree, not estimated, so `--limit` ranks by real size — and
 counted once per item however many agents a repo renders it for, so a pack that
 ships a copy in `.claude/`, `.cursor/` and twelve other dotdirs is not credited
 with fourteen skills (`scripts/measure_registry.py`).
 
-Six domains are curated end to end, each tappable on its own:
+Seven domains are curated end to end, each tappable on its own:
 **`ai`** (RAG, evals, tracing, model serving — official OpenAI, Anthropic,
 LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 **`architecture`** (DDD, clean architecture, system design, C4/ADR),
@@ -116,8 +116,13 @@ LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 TUI/dotfiles — including Impeccable, taste-skill and Huashu Design),
 **`java`** (Spring Boot, Kotlin/JVM, Quarkus, JetBrains, Camunda, Vaadin),
 **`ecommerce`** (Shopify, Magento, WooCommerce, Spree, Bagisto, PrestaShop,
-Stripe, Algolia), and **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
-networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs):
+Stripe, Algolia), **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
+networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs),
+and **`marketing`** (marketing, CRM, email campaigns and customer outreach —
+20 registries, ~2,400 measured items: cold-email and ABM sequences from
+ColdIQ, CRM setup and hygiene for Salesforce/HubSpot/Attio from LeadMagic,
+B2B RevOps lead routing, Zapier's GTM cheat codes, plus paid media, SEO/GEO
+and lifecycle email):
 
 Smaller categories fill in around them. **`efficiency`** collects the packs whose
 items exist to make an agent emit *less* — less code (`ponytail`) or fewer output

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/boost_cli/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '1.2.17'
-__version_tuple__ = version_tuple = (1, 2, 17)
+__version__ = version = '1.2.19'
+__version_tuple__ = version_tuple = (1, 2, 19)
 
 __commit_id__ = commit_id = None

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/boost_cli/data/registries.json

@@ -1,8 +1,8 @@
 {
   "generated_note": "Curated skill/rule/workflow registries for boost. est_items are research estimates; verify by tapping. list_only repos are index/awesome lists that link out.",
-  "count": 470,
-  "est_items_total": 32628,
-  "est_items_scannable": 28207,
+  "count": 486,
+  "est_items_total": 34840,
+  "est_items_scannable": 30419,
   "registries": [
     {
       "name": "Aaronontheweb/dotnet-cursor-rules",
@@ -356,6 +356,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "aaron-he-zhu/aaron-marketing-skills",
+      "url": "https://github.com/aaron-he-zhu/aaron-marketing-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Marketing staff across narrative, SEO/GEO, social, email, paid, influencer and launch, with auditor gates",
+      "est_items": 128,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "abubakarsiddik31/claude-skills-collection",
       "url": "https://github.com/abubakarsiddik31/claude-skills-collection",
@@ -404,8 +415,8 @@
       "name": "AgriciDaniel/claude-ads",
       "url": "https://github.com/AgriciDaniel/claude-ads",
       "type": "skill",
-      "category": "general",
-      "focus": "Claude-first paid-media operations skill for Claude Code across 12 ad... (34 skills, 25 workflows)",
+      "category": "marketing",
+      "focus": "Paid-media operations across 12 ad platforms: audits, budgets, creative, landing pages, attribution",
       "est_items": 59,
       "confidence": "high",
       "list_only": false,
@@ -426,9 +437,9 @@
       "name": "AgriciDaniel/claude-seo",
       "url": "https://github.com/AgriciDaniel/claude-seo",
       "type": "skill",
-      "category": "general",
-      "focus": "Universal SEO skill for Claude Code. 25 sub-skills + 18 sub-agents... (25 skills, 20 workflows)",
-      "est_items": 45,
+      "category": "marketing",
+      "focus": "Technical and content SEO: audits, backlinks, schema, clustering, local, international and ecommerce",
+      "est_items": 51,
       "confidence": "high",
       "list_only": false,
       "curated": true
@@ -1093,6 +1104,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "Cold-IQ/ColdIQ-s-GTM-Skills",
+      "url": "https://github.com/Cold-IQ/ColdIQ-s-GTM-Skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Cold email sequences, ABM messaging, buying signals, list building and deliverability",
+      "est_items": 90,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "comet-ml/opik-skills",
       "url": "https://github.com/comet-ml/opik-skills",
@@ -1141,10 +1163,10 @@
       "name": "coreyhaines31/marketingskills",
       "url": "https://github.com/coreyhaines31/marketingskills",
       "type": "skill",
-      "category": "writing",
-      "focus": "Marketing and growth-focused agent skills",
-      "est_items": 10,
-      "confidence": "med",
+      "category": "marketing",
+      "focus": "CRO, copywriting, cold email, SEO, analytics and growth-engineering skills",
+      "est_items": 50,
+      "confidence": "high",
       "list_only": false,
       "curated": true
     },
@@ -1434,6 +1456,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "ericosiu/ai-marketing-skills",
+      "url": "https://github.com/ericosiu/ai-marketing-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Growth experiments, cold outbound, sales pipeline, content ops and revenue intelligence",
+      "est_items": 27,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "Eronred/aso-skills",
       "url": "https://github.com/Eronred/aso-skills",
@@ -2094,6 +2127,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "kostja94/marketing-skills",
+      "url": "https://github.com/kostja94/marketing-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "SEO, content, 40+ page-type generators, paid ads, channel and growth strategy",
+      "est_items": 172,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "Kotlin/kotlin-agent-skills",
       "url": "https://github.com/Kotlin/kotlin-agent-skills",
@@ -2215,6 +2259,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "LeadMagic/gtm-skills",
+      "url": "https://github.com/LeadMagic/gtm-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "CRM setup and hygiene (Salesforce, HubSpot, Attio), enrichment, AI SDR, ABM and lifecycle",
+      "est_items": 206,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "lebsral/DSPy-Programming-not-prompting-LMs-skills",
       "url": "https://github.com/lebsral/DSPy-Programming-not-prompting-LMs-skills",
@@ -2259,6 +2314,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "LeoYeAI/openclaw-marketing-skills",
+      "url": "https://github.com/LeoYeAI/openclaw-marketing-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Cold email, email sequences, churn prevention, CRO, copywriting and paid ads",
+      "est_items": 39,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "lgbarn/devops-skills",
       "url": "https://github.com/lgbarn/devops-skills",
@@ -2538,9 +2604,9 @@
       "name": "minhnv0807/ai-business-skills",
       "url": "https://github.com/minhnv0807/ai-business-skills",
       "type": "skill",
-      "category": "general",
-      "focus": "63 bilingual AI marketing skills (31 VN + 31 Global) for Claude Code... (49 skills, 13 workflows)",
-      "est_items": 62,
+      "category": "marketing",
+      "focus": "Bilingual VN/global marketing SOP packs: plans, content calendars, campaign briefs, ad copy, reporting",
+      "est_items": 169,
       "confidence": "high",
       "list_only": false,
       "curated": true
@@ -2732,6 +2798,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "NEON-Rutger/B2B-revops-skills",
+      "url": "https://github.com/NEON-Rutger/B2B-revops-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "B2B RevOps: CRM migration, lead routing, deal desk, pipeline visibility, ICP and forecasting",
+      "est_items": 38,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "nexscope-ai/Amazon-Skills",
       "url": "https://github.com/nexscope-ai/Amazon-Skills",
@@ -2798,6 +2875,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "nowork-studio/notfair-plugin",
+      "url": "https://github.com/nowork-studio/notfair-plugin",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "SEO, GEO and Google/Meta ads: keyword research, audits, ad copy, landing pages, analytics",
+      "est_items": 90,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "NTCoding/claude-skillz",
       "url": "https://github.com/NTCoding/claude-skillz",
@@ -2941,6 +3029,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "OpenClaudia/openclaudia-skills",
+      "url": "https://github.com/OpenClaudia/openclaudia-skills",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "SEO, content, email, ads, analytics and growth, including Apollo outreach and backlink audits",
+      "est_items": 75,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "OpenSenseNova/SenseNova-Skills",
       "url": "https://github.com/OpenSenseNova/SenseNova-Skills",
@@ -3007,6 +3106,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "Othmane-Khadri/YALC-the-GTM-operating-system",
+      "url": "https://github.com/Othmane-Khadri/YALC-the-GTM-operating-system",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "GTM operating system: campaign strategy, multi-touch copywriting sequences, CRM and sequencer adapters",
+      "est_items": 61,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "Owl-Listener/ai-design-skills",
       "url": "https://github.com/Owl-Listener/ai-design-skills",
@@ -4250,6 +4360,17 @@
       "list_only": true,
       "curated": true
     },
+    {
+      "name": "zapier/gtm-cheat-codes",
+      "url": "https://github.com/zapier/gtm-cheat-codes",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "GTM field guide: campaign planning, cross-CRM opportunity sync, customer proof, lead stewardship",
+      "est_items": 21,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "zhaoxuya520/reverse-skill",
       "url": "https://github.com/zhaoxuya520/reverse-skill",
@@ -4283,6 +4404,39 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "zubair-trabzada/ai-marketing-claude",
+      "url": "https://github.com/zubair-trabzada/ai-marketing-claude",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Site audit through to email sequences, ad campaigns, content calendars and client-ready reports",
+      "est_items": 20,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
+    {
+      "name": "zubair-trabzada/ai-sales-team-claude",
+      "url": "https://github.com/zubair-trabzada/ai-sales-team-claude",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "Sales outreach workflow: ICP, prospecting, qualification, objection handling, follow-up",
+      "est_items": 19,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
+    {
+      "name": "zubair-trabzada/geo-seo-claude",
+      "url": "https://github.com/zubair-trabzada/geo-seo-claude",
+      "type": "skill",
+      "category": "marketing",
+      "focus": "GEO / AI-search optimisation: citability scoring, AI-crawler analysis, schema, brand mentions",
+      "est_items": 21,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "0xfurai/claude-code-subagents",
       "url": "https://github.com/0xfurai/claude-code-subagents",
@@ -4305,6 +4459,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "aitytech/agentkits-marketing",
+      "url": "https://github.com/aitytech/agentkits-marketing",
+      "type": "workflow",
+      "category": "marketing",
+      "focus": "Marketing-automation commands and agents: campaigns, ads, analytics, content ops",
+      "est_items": 399,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "aiwonglab/claude_code_agents",
       "url": "https://github.com/aiwonglab/claude_code_agents",
@@ -4756,6 +4921,17 @@
       "list_only": false,
       "curated": true
     },
+    {
+      "name": "gtmagents/gtm-agents",
+      "url": "https://github.com/gtmagents/gtm-agents",
+      "type": "workflow",
+      "category": "marketing",
+      "focus": "GTM agents and commands for sales, marketing, customer success and revenue operations",
+      "est_items": 653,
+      "confidence": "high",
+      "list_only": false,
+      "curated": true
+    },
     {
       "name": "hesreallyhim/a-list-of-claude-code-agents",
       "url": "https://github.com/hesreallyhim/a-list-of-claude-code-agents",

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/boost_skill_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: boost-skill-cli
-Version: 1.2.17
+Version: 1.2.19
 Summary: boost — Homebrew for AI coding skills
 Author: Jonathan Reyes
 License: GNU GENERAL PUBLIC LICENSE
@@ -818,15 +818,15 @@ rather than failing: no extra, no key, or no built store each fall back to BM25,
 and `boost doctor` reports which engine is actually serving and why.
 
 Want the whole ecosystem instead of the starter set? boost ships a **curated
-registry catalog** — 460+ classified GitHub registries of skills, Cursor/Windsurf
+registry catalog** — 480+ classified GitHub registries of skills, Cursor/Windsurf
 **rules**, and Claude Code **workflows** (slash commands & subagents),
-collectively indexing ~28,000 scannable items. Every `est_items` is *measured*
+collectively indexing ~30,000 scannable items. Every `est_items` is *measured*
 from the repo's file tree, not estimated, so `--limit` ranks by real size — and
 counted once per item however many agents a repo renders it for, so a pack that
 ships a copy in `.claude/`, `.cursor/` and twelve other dotdirs is not credited
 with fourteen skills (`scripts/measure_registry.py`).
 
-Six domains are curated end to end, each tappable on its own:
+Seven domains are curated end to end, each tappable on its own:
 **`ai`** (RAG, evals, tracing, model serving — official OpenAI, Anthropic,
 LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 **`architecture`** (DDD, clean architecture, system design, C4/ADR),
@@ -834,8 +834,13 @@ LangChain, Qdrant, Elastic, MLflow, Langfuse, Arize, W&B, vLLM libraries),
 TUI/dotfiles — including Impeccable, taste-skill and Huashu Design),
 **`java`** (Spring Boot, Kotlin/JVM, Quarkus, JetBrains, Camunda, Vaadin),
 **`ecommerce`** (Shopify, Magento, WooCommerce, Spree, Bagisto, PrestaShop,
-Stripe, Algolia), and **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
-networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs):
+Stripe, Algolia), **`infra`** (Docker, Kubernetes, OpenShift, Helm, GitOps,
+networking — plus official Azure, Grafana, Pulumi, Flux, DigitalOcean packs),
+and **`marketing`** (marketing, CRM, email campaigns and customer outreach —
+20 registries, ~2,400 measured items: cold-email and ABM sequences from
+ColdIQ, CRM setup and hygiene for Salesforce/HubSpot/Attio from LeadMagic,
+B2B RevOps lead routing, Zapier's GTM cheat codes, plus paid media, SEO/GEO
+and lifecycle email):
 
 Smaller categories fill in around them. **`efficiency`** collects the packs whose
 items exist to make an agent emit *less* — less code (`ponytail`) or fewer output

{boost_skill_cli-1.2.17 → boost_skill_cli-1.2.19}/boost_skill_cli.egg-info/SOURCES.txt

@@ -303,6 +303,7 @@ docs/roadmap/items/make-lint-masks-actionlint-failures.md
 docs/roadmap/items/make-the-roadmaps-discoverable.md
 docs/roadmap/items/markdown-consistency-markdownlint-cli2.md
 docs/roadmap/items/markdownlint-lints-the-fuzz-corpus.md
+docs/roadmap/items/marketing-crm-domain-had-no-category.md
 docs/roadmap/items/mcp-already-covered-defeater.md
 docs/roadmap/items/mcp-aware-skills.md
 docs/roadmap/items/mcp-check-skills-before-starting-a-task.md

boost_skill_cli-1.2.19/docs/roadmap/items/marketing-crm-domain-had-no-category.md

@@ -0,0 +1,73 @@
+---
+id: marketing-crm-domain-had-no-category
+board: code
+section: internals
+status: inflight
+category: Catalog · Curation
+complexity: M
+impact: High
+wow: 3
+note: the four marketing registries already carried were filed under writing and general, so the category returned nothing
+order: 124
+owner: loop/marketing-crm-registries
+pr: 551
+title: <code>--category marketing</code> matched nothing, while four marketing registries sat in the catalog under other names
+---
+<b>The catalog carried marketing registries and could not hand you one.</b>
+<code>boost tap --catalog --category marketing</code> returned zero rows, because no row was filed
+that way. Four already existed: <code>coreyhaines31/marketingskills</code> (45,374 stars) under
+<code>writing</code>, and <code>minhnv0807/ai-business-skills</code>,
+<code>AgriciDaniel/claude-ads</code> and <code>AgriciDaniel/claude-seo</code> under
+<code>general</code>. Their items are named <code>cold-email</code>, <code>email-sequence</code>,
+<code>market-emails</code>, <code>ads-attribution</code>, <code>seo-backlinks</code>. Nothing about
+that is writing, and nothing about it is general.
+
+<b>The misfiling has one cause, and it is the counterpart of the <code>ui</code> trap.</b>
+<code>ui</code> is pinned because a scorer keyed on the <em>repo name</em> files
+<code>ai-design-skills</code> as design. This domain fails one step earlier: every repo in it
+describes itself in agent vocabulary before it says what the skills do — &ldquo;345 skills for
+Claude Code, Codex, Gemini CLI, Cursor and 8 more coding agents&rdquo; — so a scorer keyed on the
+<em>README</em> lands on <code>general</code> every time and never reaches the word
+<code>hubspot</code>. Reading item names is what separates them, which is the rule
+<code>CLAUDE.md</code> already states and which this batch is a second proof of.
+
+<b>Twenty registries, 2,388 measured items, seventh curated domain.</b> Sixteen new rows plus the
+four recategorised. Every count is <code>scripts/measure_registry.py</code> against a fresh
+<code>--filter=blob:none --depth=1</code> clone, never the repo's advertised figure — and the two
+recategorised counts moved a long way when re-measured: <code>marketingskills</code> from 10 to
+<b>50</b> items and <code>ai-business-skills</code> from 62 to <b>169</b>. Both old numbers were
+that repo's own README total from an earlier release. Moving a row's category is not a reason to
+trust the number attached to it, so <code>MARKETING_MEASURED</code> pins both.
+
+<b>The cut that nearly shipped a CRM category with no CRM in it.</b> Thirty candidates were cloned
+and measured; ten were dropped — <code>mikiarlo3/awesome-growth-hacking-skills</code> scans to
+<b>0</b> items, <code>CosmoBlk/email-marketing-bible</code> to 1,
+<code>mysticaltech/marketingskills</code> is a stale fork of a repo already in the batch, and
+<code>gtm-skills/gtm</code> ships 5 items under names (<code>rep</code>, <code>scout</code>,
+<code>closer</code>) that say nothing. Ranking the survivors by adoption then produced a top 20 in
+which <b>the CRM half of the domain's own name had no coverage at all</b>: the CRM registries are
+the least-starred rows in the batch — <code>LeadMagic/gtm-skills</code> at 42 stars for 206 items
+covering Salesforce, HubSpot and Attio setup, <code>NEON-Rutger/B2B-revops-skills</code> at 45 for
+CRM migration and lead routing — against 45,374 for the top content-marketing pack. Popularity and
+coverage disagree here, and coverage wins:
+<code>TestMarketingDomain::test_every_advertised_subdomain_has_a_registry</code> names the registry
+carrying each of the four advertised sub-domains (CRM, RevOps, cold outreach, campaigns) and fails
+if one is dropped or flagged <code>list_only</code>.
+
+<b>Discovery was <code>boost_search</code> first, and it earned its ten seconds by narrowing the
+question.</b> The corpus already tapped on this machine returned <code>cold-email</code>,
+<code>deal-outreach</code>, <code>ai-cold-outreach</code> and <code>emails</code> — real hits, from
+repos mostly already in the catalog. That is what said the gap was not <em>items</em> but a
+<em>category</em>: the skills existed and were unreachable by domain. GitHub code search over
+<code>filename:SKILL.md</code> with CRM vocabulary (<code>hubspot</code>,
+<code>salesforce</code>, <code>pipeline</code>) is what surfaced the two CRM registries; repo search
+alone never returned either, because neither repo's description mentions a CRM by name.
+
+<b>What is deliberately not here.</b> Six live candidates were measured and left out for being
+mixed rather than bad: <code>Varnan-Tech/opendirectory</code> (64 items, but
+<code>dependency-update-bot</code> and <code>explain-this-pr</code> sit beside the marketing ones),
+<code>TheCraigHewitt/skills</code> (65 items, roughly a quarter marketing, the rest founder ops),
+<code>citedy/adclaw</code>, <code>manojbajaj95/claude-gtm-plugin</code>,
+<code>markster-public/markster-os</code> and <code>hyperfx-ai/marketing-skills</code>. They are
+recorded here rather than merely dropped, so the next sweep does not re-derive them from the same
+search and re-litigate the same call.