Skip to content

scope.md

Latest commit

 

History

History
52 lines (40 loc) · 2.31 KB

scope.md

File metadata and controls

52 lines (40 loc) · 2.31 KB

ai-seo-magic-button — scope & positioning (AL-722)

Promise

Point it at your site, get a whole-site AEO/GEO audit plus a ready-to-run plan your agent can execute.

The output is an analysis and a plan, not direct edits. The product audits and proposes; your agent (in any MCP host) executes. This "analyze and propose, not fix" framing is the settled positioning — the apply helpers exist to make the plan trivially executable, but they run the plan you can read, never silent edits to a live site.

What it is

  • A whole-site AEO/GEO/LLM-visibility auditor. Crawls the sitemap, audits every sampled page, and layers in citation-gap signal.
  • A generator of a portable, structured plan (plan.json + a markdown checklist) where every item is self-contained and agent-executable.
  • A thin orchestration layer over two engines we already ship:
    • ai-seo (AI Citation Toolkit) — audit, score, rewrite for AEO/GEO.
    • citation-intelligence — which URLs AI engines cite, and gaps.
  • Shipped on all three surfaces off one shared core lib: a Claude skill (the magic-button UX), a Claude plugin (bundles the skill), and an MCP server (generate_seo_plan, runs in Cursor/Cline/Windsurf/Claude Desktop).

What it is not

  • Not a single-URL checker — that is the existing web tool / GitHub Action. The value here is whole-site coverage and a runnable plan.
  • Not a tool that edits your live site directly. It writes a plan; execution is explicit and reviewable.
  • Not a re-implementation of the audit engines. It drives them.
  • Not dependent on any API keys for the core audit→plan flow (the ai-seo and citation-intelligence audit/predict paths are rule-based). LLM-dependent rewrites degrade to prompt templates the executing agent finishes.

Dependencies

  • @automatelab/ai-seo-mcp
  • @automatelab/citation-intelligence

Both are spawned as stdio subprocesses by the core lib. Resolve via npx by default, or point AISEO_MCP_PATH / CITATION_MCP_PATH at local builds.

Target audience

  • Site owners and content teams who want AI-citation visibility (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews) and want a plan their coding agent can just run.
  • Agencies running client sites who want a repeatable audit → plan → execute → verify loop.