Core Skill

One Prompt, the Right Skill — How Dispatch Routes Your Workflow

Stop managing 36 skills manually. Dispatch reads your request, selects the right skill combination, routes to the right data endpoints, and returns a synthesized answer in one prompt.

Get on GitHub

The Dispatch Skill: One Prompt. The Right Answer. Every Time.

You’ve got the tools. You’ve got the data. You’ve got the knowledge. What you don’t have is a fast path from “I need to understand why this business isn’t ranking” to an actual answer.

Right now that path runs through: open LocalSEOData, run the query, open the spreadsheet, cross-reference the last audit, pull competitor profiles one at a time, try to hold the framework in your head while switching tabs, synthesize it all while trying to remember which window had the completeness score.

The work isn’t hard. The friction is the problem.

You know which data you need. You know which analysis framework applies. But translating “why isn’t this business ranking” into a sequence of tool calls, skill invocations, and synthesis steps takes time — and that time multiplies across every client, every audit, every question.

What Changes When Dispatch Is Running

You describe what you need in plain language. Dispatch figures out the rest.

“Why isn’t this business in the top 3 for [keyword] in [city]?” — that’s the complete input.

Dispatch reads the intent, selects the right combination of strategy skills and data tools, orchestrates the workflow, and returns a synthesized answer. You didn’t have to know which skill to call. You didn’t have to know which endpoint to query. You just said what you needed.

The dispatch skill sits between your natural language request and the 36 skills in the LocalSEOSkills library. It handles the routing decisions that would otherwise require you to know:

  • Which of the 24 strategy skills applies to this request
  • Which of the 12 tool skills need to fire
  • Which LocalSEOData endpoints return the relevant data
  • What order to call them in
  • How to synthesize the outputs into a coherent answer

You don’t manage any of that. Dispatch does.

See It Work: One Prompt, Complete Analysis

Prompt: "New client — Smith's HVAC, 4521 Main Street Dallas TX.
Not ranking for 'HVAC repair Dallas'. Why not and what are the top 5 fixes?"

Skills that fire: dispatch → gbp-optimization + local-competitor-analysis + localseodata-tool
Data pulled: business_profile, profile_health, local_pack ("HVAC repair Dallas"),
             competitor_gap, business_profile ×3 (top competitors)

DISPATCH ANALYSIS: Smith's HVAC

WHY YOU'RE NOT RANKING
Primary gap: GBP completeness 61% vs. competitor average 89%
Category gap: Missing 5 secondary categories all top 3 competitors claim
Review gap: 34 reviews at 4.6★ vs. competitor average 156 reviews at 4.8★
Response rate: 14% vs. competitor average 87%

TOP 5 FIXES (priority order)
1. Add secondary categories: "Air Conditioning Contractor", "Furnace Repair
   Service", "HVAC Installation", "Indoor Air Quality Service",
   "Emergency HVAC Service" — 15 minutes, highest leverage
2. Respond to all 29 unanswered reviews — 2 hours, direct ranking signal
3. Add attributes: "24/7 availability", "Emergency service",
   "Licensed", "Insured" — 5 minutes
4. Upload 40+ photos (current: 6, competitor avg: 47) — 1 hour
5. Launch review generation campaign — 90-day priority

Estimated time to top-5 with consistent execution: 60–90 days

One prompt. 45 seconds. A prioritized answer your client can act on today.

What happened in those 45 seconds:

  1. Dispatch parsed your request and identified task type: ranking diagnosis
  2. Selected the relevant skills: gbp-optimization for profile analysis, local-competitor-analysis for benchmarking
  3. Selected the data tool: localseodata-tool
  4. Determined which endpoints: business_profile, profile_health, local_pack, competitor_gap
  5. Orchestrated the calls in the correct sequence
  6. Synthesized the results into a prioritized action list
  7. Returned everything in plain language with time estimates

You typed one sentence. Dispatch handled the orchestration.

What the Dispatch Skill Does

Reads natural-language local SEO requests and identifies task type, required skills, and required data. Whether you ask “audit this GBP,” “why are we losing to [competitor],” “what’s our AI visibility,” or “create a proposal for this prospect” — dispatch recognizes the task type and routes accordingly.

Routes to single skills or multi-skill combinations based on task complexity. A GBP audit might need only gbp-optimization. A full ranking diagnosis needs gbp-optimization + local-competitor-analysis + geogrid-analysis. Dispatch handles both.

Handles ambiguous requests with a clarifying question before proceeding. If you ask “analyze this business” without specifying the analysis type, dispatch asks: “What specifically do you want to analyze? GBP completeness? Competitive position? AI visibility? Citation status?” One clarification, then it routes.

Orchestrates sequential skill chains. Complex workflows like audit → proposal → deliverable require multiple skills firing in sequence, with the output of one feeding the input of the next. Dispatch manages the chain.

Works with all 24 strategy skills and all 12 tool skills. The full LocalSEOSkills library is routable through dispatch — GBP optimization, competitive analysis, citations, reviews, rankings, AI visibility, content, reporting, scale operations, and all tool integrations.

Selects data tools based on what’s connected and what the task requires. If LocalSEOData is connected, it’s the default for most workflows. If Local Falcon is connected and you ask about ranking trends, dispatch routes there. If you ask about deep backlinks and Ahrefs is connected, dispatch routes there. The routing is context-aware.

Draws on CLAUDE.md and tool-routing.md for routing context. These configuration files tell dispatch which tools are available, which endpoints exist, and which skill combinations work together. Dispatch reads them on every request.

Maintains session context so follow-up prompts route correctly. After running an audit, you can say “now turn that into a proposal” — dispatch remembers the audit output and routes to client-deliverables with the prior context intact.

Who Uses Dispatch and How

The Freelancer in a New Client Session

90 minutes before the onboarding call. You need an audit, competitive analysis, and draft proposal. Three prompts:

"Audit the GBP for [Business Name] at [Address] targeting [keyword] in [city]."
"Compare them to the current top 3 in the local pack."
"Turn this into a 6-month engagement proposal at $1,200/month."

Dispatch handles the routing each time. The gbp-optimization skill, then local-competitor-analysis, then client-deliverables. You have a complete proposal before the call starts.

The Agency Running Monthly Maintenance

15 clients. One prompt per client to identify which need attention:

"Quick health check for [Client Name] — any urgent issues this month?"

Dispatch routes to the appropriate audit endpoints, pulls the data, and flags what needs attention. Portfolio health check across 15 clients in 30 minutes instead of 3 hours.

The Business Owner Doing This for the First Time

Doesn’t know which skill to call. Doesn’t know what data they need. Just knows their business isn’t showing up when people search.

"My business is [Name] at [Address]. We're not showing up for [service] in [city]. What's wrong?"

Dispatch is why they don’t have to know the skill library. Describe the problem, get the answer. The routing complexity is invisible.

The Practitioner Hitting an Edge Case

Something outside a standard workflow. A client asking about a specific situation that doesn’t fit the usual templates.

"A competitor's GBP shows a business name with keywords stuffed in.
How do I report this and what's the likely outcome?"

Dispatch identifies the closest skill combination (gbp-optimization for context, potentially gbp-suspension-recovery for policy knowledge), flags that it’s an edge case, and asks if you want it to proceed with that routing.

What 36 Skills Without Dispatch Looks Like

Without dispatch, LocalSEOSkills is 36 separate files. Useful — but manual.

You have to know which skill applies to your request. You have to call it explicitly by referencing the skill name or invoking it directly. You have to manage the workflow yourself when multiple skills are needed. You have to know which data endpoints to request.

It works. Practitioners who know the skill library well can navigate it directly. But every prompt requires you to think about the routing before you think about the actual question.

With dispatch, the library becomes a single intelligent interface. Same difference as knowing where every tool in a workshop is vs. having a master craftsman hand you the right one before you ask.

Without dispatch:

"Using the gbp-optimization skill, audit [Business Name].
Then using local-competitor-analysis, compare to [Competitor A], [Competitor B], and [Competitor C].
Then using client-deliverables, format this as a proposal."

With dispatch:

"Audit [Business Name], compare to competitors, and give me a proposal draft."

Same output. Different friction. The difference compounds across dozens of prompts per day.

Get This Skill — It’s Free and Open Source

Dispatch is part of the LocalSEOSkills library. MIT licensed. Free to use, free to modify, free to inspect.

The skill file is readable — you can open dispatch.md and see exactly how it determines routing, which keywords trigger which skills, and how it handles ambiguity. There’s no black box. The routing logic is transparent.

Installation:

  1. Download LocalSEOSkills from GitHub
  2. Upload to Claude.ai (zip file) or configure in Claude Code (directory reference)
  3. Dispatch is active immediately — no additional configuration

First prompt to verify:

"I have a new client, [Business Name] at [Address], targeting [keyword] in [city].
Give me a complete audit with competitive comparison and quick wins."

If dispatch is working, you’ll see it identify the task type, select the skills, pull the data, and return a synthesized answer. One prompt, complete output.

The routing happens automatically. The synthesis happens automatically. Your job is to describe what you need. Dispatch handles how to get it.

What Dispatch Routes To

Audit and Analysis Skills:

  • local-seo-audit — comprehensive 8-dimension audits
  • gbp-optimization — GBP profile analysis and recommendations
  • local-competitor-analysis — competitive benchmarking
  • geogrid-analysis — geographic ranking visualization
  • ai-local-search — AI platform visibility measurement

Content and Deliverable Skills:

  • client-deliverables — proposals, reports, client-ready documents
  • local-landing-pages — location page generation
  • local-schema — LocalBusiness JSON-LD generation
  • gbp-posts — GBP post calendar creation

Data and Monitoring Skills:

  • review-management — review response and analysis
  • local-citations — citation audit and strategy
  • local-reporting — monthly performance reports
  • multi-location-seo — portfolio-level analysis

Tool Skills:

  • localseodata-tool — primary data layer (36 endpoints)
  • local-falcon-tool — geogrid trends and Falcon Guard
  • lsa-spy-tool — LSA market intelligence
  • And 9 more specialist integrations

Every skill is routable through dispatch. The question is what you need — dispatch determines which skills answer it.

Skill Documentation

For technical details on how this skill works, what data it pulls, and complete prompt reference, see the full skill documentation.

All 36 skills. Free. Open source. Get on GitHub →