gooseworks-ai/goose-skills / skills/marketing/improve-search-visibility
improve-search-visibility
VerifiedSEO and AI-answer visibility. Use for a site SEO or technical audit with ranked fixes, competitor keyword and topic gaps, a topic cluster, SEO metrics for domains, a programmatic SEO page-set plan, or checking where ChatGPT and Google name the brand. Delivers the requested audit, comparison, map or blueprint with dated evidence and clear data limits. Also covers app store listings and local search. Drafts changes; never edits live. Full pages go to make-content; paid search goes to run-paid-campaigns.
references/entity-and-local.md
View on GitHub# Entity footprint and local
Load at step 1 (free pass, atom entity-footprint).
## The canonical facts (set once, from the brain; the human confirms)
Name and legal name · one-line description (the canonical sentence) · primary category · founding year and founders · HQ city · canonical domain · logo · pricing model · for local: name, address, phone (NAP) and hours per location. The most common cause of a wrong brand description in AI answers is an old positioning left on one profile: fix those first.
## How the name is said and misheard
Read any pronunciation fact in the brand's learnings (`gooseworks:brand_knowledge_search` for learnings when `brand_knowledge_*` tools are listed, else `brand_read`; written `Pronounce "<term>" as "<spoken form>"`). With none, work from the written name. People type what they heard: from the spoken form, list the spellings a buyer would try ("Goose Works", "Goosework", "Gooseworx") and use them in brand questions, in the Organization markup's `alternateName`, and in the "who we are" line when a namesake wins. Voice assistants read names the same way. Never write a pronunciation fact here — this playbook doesn't record brand facts; if a name is easy to misread and no fact exists, say so in "couldn't check" as a question for the brand.
## Profiles that matter, by type (check the top five for this brand)
| Type | Must-have | Nice-to-have |
|---|---|---|
| All | own About page with the facts, Organization markup with sameAs, LinkedIn page, Crunchbase, Wikidata (only if notable) | YouTube, X/Instagram |
| B2B SaaS | G2, Capterra/GetApp/Software Advice (one account), TrustRadius, integration-partner directories, public docs and pricing pages | Product Hunt, AlternativeTo, SaaSworthy, GitHub |
| Store (DTC) | Trustpilot (claimed), Google Merchant Center store rating, Amazon brand store if selling there, retailer listings | Product Hunt for launches, disclosed founder presence on Reddit |
| Local / physical | Google Business Profile, Apple Business Connect, Bing Places, Yelp, Facebook page, the industry directory (Healthgrades, Zocdoc, Avvo, Houzz, Tripadvisor) | Nextdoor, chamber of commerce |
| Consumer app | App Store and Play listings, developer site with support and privacy pages, Product Hunt | AlternativeTo, Reddit, YouTube demos |
| Services | Clutch-type directory for the niche, LinkedIn (firm + founder), Google Business Profile if clients visit | podcast/guest pages (gap only) |
## Scoring each profile
complete / inconsistent / missing / unclaimed, against: same name; the canonical one-liner (word for word or as a clear prefix); same primary category in that site's taxonomy; same founding facts; same domain; same logo; same pricing model or no price rather than an old one; same NAP (local).
## Reading profiles
Read public profiles with the host's fetch where the site allows it. G2, Capterra, Trustpilot and Amazon forbid automated collection: never fetch them — ask the human for a screenshot or their vendor-portal view, or mark "not checked". Crunchbase and app stores often block fetchers — same fallback.
## The fix is field values, not advice
Write the exact value per field (see `finished-fix-examples.md` §d) and how to edit it (claim the profile, vendor portal, suggest-an-edit). Unclaimed profiles: the claim steps.
## Local module (any business with in-person customers)
Google Business Profile is only for businesses that meet customers in person or at their location. Per location: business name exactly as on the signage (no keywords added); the single most specific primary category; 2–4 secondary; hours incl. holiday overrides; address or service area; phone; website with UTM; services/products; description (≤ 750 characters, facts first); photos; review count, average, recency (last 30/90 days) and reply rate. NAP identical across the profile, the site, Apple, Bing, Yelp and the industry directory. Sample local questions per city.
Review gaps → a review-request flow brief for email-lifecycle (every customer, the direct review link, timing after the visit). Replying to reviews is not covered yet (manage-reviews, backlog): report the reply rate and say so plainly. Never fake reviews, never an incentive tied to a positive rating or left undisclosed, never asking only happy customers; a platform's own disclosed review programme is fine.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |