Loading ToolBench
Agent-first SEO quality, intent and opportunity engine with CLI and optional MCP server.
Seven tools with complete input schemas and descriptions, but significant quality gaps. All tools are READ_ONLY (good for safety), but descriptions are generic and lack LLM-optimized guidance. Parameter descriptions are minimal (avg ~40 chars vs baseline 72). No output schemas documented. Tool names follow verb_noun pattern but lack specificity about when to use each tool vs alternatives. No error handling guidance, no recovery paths, no actionable next steps in descriptions.
Checks image alt attribute coverage in HTML, returning total images, count with alt, missing alt count, coverage ratio, and missing tag snippets.
Returns the connection status and configured external services (Gemini, GA4, Google Search Console).
Analyzes search intent for a given keyword, returning primary intent, secondary intent, confidence scores, detected signals, and recommendations.
Returns the agent manifest with project metadata, supported clients, standard tools, and recommended configuration.
Prioritizes SEO opportunities based on impressions, clicks, position, conversions, and commercial intent, returning a final score, priority level, score breakdown, and recommended action.
Returns privacy audit information including secrets handling, local files ignored, external services, data boundaries, and safety rules.
No output schemas documented. LLMs cannot plan downstream calls or extract required fields. All 7 tools lack return type documentation.
Parameter descriptions are minimal and lack actionable guidance. E.g., 'The keyword to analyze for search intent' (40 chars) vs baseline 72 chars. No format constraints, no examples of valid input, no dependency hints.
Tool descriptions lack LLM-optimized guidance on WHEN to use each tool. No distinction between agent_seo_manifest, agent_seo_connection_status, and agent_seo_privacy_audit, all are discovery tools but lack clear sequencing hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2026-07-28+ | v2 |
Scores markdown content for SEO quality, returning overall score, grade, publishing readiness, detailed checks, metrics, and recommendations.
No error handling or recovery guidance. Tools return dicts but no documentation of error cases, failure modes, or what LLM should do if a call fails (retry, ask user, abort).
Mutually exclusive parameters not documented. agent_seo_score_content accepts both 'markdown' and 'file_path'; agent_seo_check_image_alt accepts both 'html' and 'file_path'. No description states which takes precedence or what happens if both are provided.