keyword-research

This skill turns seed topics into a prioritized keyword opportunity set using OpenSEO MCP data, helping the user decide what to target, what to save, and what to research next. It addresses the common SEO problem of chasing high-volume terms without regard to intent, difficulty, or a site's real ability to rank, by grounding recommendations in first-party demand and objective metrics.

The workflow normalizes seeds into distinct research angles and, when Search Console is connected, starts from real first-party demand, pulling performance data, filtering to striking-distance positions (roughly 5-20) client-side, and hydrating those queries with metrics. It orchestrates a set of read-oriented OpenSEO MCP tools: research_keywords for discovery, get_keyword_metrics to attach volume, keyword difficulty, intent, CPC, and trends, get_ranked_keywords for a target domain or page, get_serp_results to inspect intent, and local-SEO tools for location-based topics. It prioritizes by product fit, clear intent, reasonable difficulty, useful volume/CPC signal, and competitive plausibility, then presents a shortlist plus an opportunity table and next actions. Guardrails require it to not invent metrics (writing 'unknown' when data is absent) and to save keywords only after explicit user confirmation.

It is aimed at SEO practitioners, content strategists, and marketers using the OpenSEO platform. It requires a projectId and one or more seeds, and is a read-and-recommend workflow whose only write action, save_keywords, is gated behind explicit confirmation.