MCP server for SEOAgent — 24 tools for AI agents to perform keyword research, rank tracking, site audits, competitor analysis, backlink intelligence, and SEO strategy generation
Four tools evaluated; all have basic descriptions (50-80 chars) and input schemas with types. However, parameter descriptions are minimal or absent, output schemas are undocumented, and error handling guidance is missing. Tool names follow verb_noun convention but lack clarity on distinctions (e.g., 'domain_reputation' vs 'domain_reputation_history'). No evidence of pagination, result limits, or recovery guidance. Schemas present but sparse, parameters lack constraints (enums, ranges, format hints). This is a typical community server with functional but incomplete definitions.
Find sites that link to competitors but not to you — backlink acquisition opportunities
Get backlink profile: top referring domains, anchor text distribution, dofollow/nofollow ratio
Get domain reputation: DR score, referring domains count, and backlink profile summary
Track domain reputation (DR) changes over time
Parameter descriptions missing or trivial. 'domain' param has description 'Domain to check (defaults to project domain)' but no format hint (FQDN? subdomain allowed?). 'project' param relies on env var fallback but does not explain what a project slug is or how to discover valid values.
Output schemas undocumented. No visible schema definition for return values. LLMs cannot plan downstream calls or extract fields without knowing what 'domain_reputation' returns (is it a number 0-100? object with DR, referring_domains, backlink_profile?). Blocks tool chaining.
No pagination or result limits documented. 'seoagent_backlink_profile' accepts 'limit' param but no description of default, max, or behavior when limit exceeds available data. 'seoagent_backlink_opportunities' similarly lacks guidance on result size and pagination.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | 2026-07-28+ | v2 |
No error handling or recovery guidance. Tool descriptions do not explain failure modes (e.g., invalid domain, missing project, API rate limit, auth failure). LLMs have no guidance on retry strategy or fallback actions.
Parameter constraints missing. 'days' param in 'seoagent_domain_reputation_history' has no min/max (is 1 day valid? 365? 10000?). 'limit' params lack bounds. Free-form numeric inputs invite hallucinated values.