Search Engine Launch (search-engine-launch)

A universal, technology-agnostic Agent Skill designed to safely prepare, audit, launch, and troubleshoot any public website or web application for search-engine discovery, crawlability, indexing, search identity, and webmaster-tool operations.

What This Skill Is

search-engine-launch establishes an engineering-grade standard for search engine launch workflows. Instead of prescribing framework-specific hacks or assuming specific tooling, it grounds search-engine readiness in public web standards (HTTP status codes, <head> metadata, canonical linking, robots.txt, XML sitemaps, and Schema.org structured data).

It adapts dynamically to whatever stack is detected—including static HTML, Next.js, React, Astro, Vue, Nuxt, SvelteKit, Laravel, Django, Flask, FastAPI, Express, Ruby on Rails, ASP.NET, Go, Spring, or headless CMS architectures.

Installation & Setup

You can install this skill globally or within a specific repository across any supported AI agent harness:

1. Cross-Runtime Global Directory (~/.agents/skills/)

git clone https://github.com/naphiertech/search-engine-launch.git ~/.agents/skills/search-engine-launch

2. Antigravity Global Directory (~/.gemini/config/skills/)

git clone https://github.com/naphiertech/search-engine-launch.git ~/.gemini/config/skills/search-engine-launch

3. Project-Specific Workspace (.agents/skills/)

For team repos, add it directly inside your project repository:

git clone https://github.com/naphiertech/search-engine-launch.git .agents/skills/search-engine-launch

Problems This Skill Prevents

  • Site Name & Identity Mismatches: Prevents Google from displaying hosting providers (e.g., "Vercel", "Heroku") or random domain strings instead of the true brand name.
  • The alternateName Hallucination Trap: Strictly forbids AI agents from automatically inventing alternate site names, acronyms, or stuffing domains into Schema.org alternateName.
  • Structured Data Spam: Blocks the fabrication of fake 5-star ratings, non-existent review counts, unverified credentials, and artificial metrics that trigger search penalties.
  • Favicon Blind Spots: Distinguishes between local file presence and live public crawlability (HTTP 200, valid MIME types, square dimensions, and robots access).
  • Crawl & Quota Abuse: Halts endless "Request Indexing" loops, automated quota evasion, and redundant sitemap resubmissions.
  • Premature Panic-Editing: Enforces the STOP-CHANGING discipline to prevent churn while search engines process updates.
  • Soft 404s & Duplicate Indexation: Ensures missing routes return real HTTP 404s and parameterized filter queries do not dilute crawl equity.
  • Robots.txt Security Fallacies: Stops developers from using robots.txt or indexing tags as a fake security measure for private routes.

The Universal Workflow

AUDIT → DEFINE IDENTITY → IMPLEMENT → VERIFY LOCAL BUILD → DEPLOY → VERIFY LIVE OUTPUT → SEARCH ENGINE TOOLS → STOP & OBSERVE
  1. Audit & Discovery: Inspect project stack, routing, and verify whether the production origin is confirmed.
  2. Define Identity: Explicitly decouple Site Name, Page Title, Canonical URL, Favicon, and Snippet description.
  3. Standards-First Implementation: Implement clean <head> metadata, self-canonical tags, sitemaps, robots.txt, and JSON-LD.
  4. Pre-Deploy Verification Gate: Validate local build outputs and syntax.
  5. Post-Deploy Live Verification Gate: Verify live HTTP 200/404 headers, MIME types, and crawler accessibility.
  6. Search Engine Webmaster Launch: Setup Google Search Console and Bing Webmaster Tools without quota exhaustion.
  7. Stop & Observe: Classify subsequent observations and respect search engine indexing latency.

Example Trigger Requests

You can invoke this skill using natural prompts such as:

  • "Use search-engine-launch and prepare this site for Google."
  • "Audit this project for Search Console launch."
  • "Use search-engine-launch to fix indexing and sitemap issues."
  • "Prepare this website for Google and Bing."
  • "Set up structured data and site name verification for our production domain."
  • "Audit our production deployment for soft 404s, favicon crawlability, and canonical hygiene."

What This Skill Does NOT Guarantee

  • No Guaranteed Search Rankings: Technical SEO optimizes eligibility, crawl hygiene, and identity clarity; search engines make independent ranking decisions.
  • No Instant Indexing: Search engine crawl queues and indexing pipelines operate asynchronously.
  • No Guaranteed Custom Snippets: Google and Bing dynamically generate snippets based on search query relevance.
  • No Guaranteed Site Name or Favicon Display: Search engine algorithms make the final automated selection from on-page and domain-level signals.

Package Structure

search-engine-launch/
├── SKILL.md                              # Core principles, rules, quick reference, red flags
├── README.md                             # Universal documentation and invocation guide
├── LICENSE                               # MIT License
├── references/
│   ├── audit-workflow.md                 # Full stack discovery & verification protocol
│   ├── search-identity.md                # Site name hierarchy, alternateName rules, favicons
│   ├── crawl-indexing.md                 # Sitemaps, robots.txt, canonicals, soft-404 prevention
│   ├── metadata-structured-data.md       # Meta tags, Open Graph, Schema.org anti-spam rules
│   ├── framework-adaptation.md           # Stack reasoning (Next.js, Laravel, Django, Astro, etc.)
│   ├── google-search-console.md          # GSC verification, Live Test vs Index, quota rules
│   ├── bing-webmaster-tools.md           # Multi-engine setup, GSC import, IndexNow
│   └── verification-checklist.md         # Pre-flight and post-deploy execution gates
└── evals/
    ├── evals.json                        # 15 standardized pressure evaluation scenarios
    ├── eval_runner.py                    # Static validator and suite runner
    └── run_full_eval_suite.py            # Comprehensive 15-scenario test runner