Loading ToolBench
OSP Marketing Tools using Model Context Protocol
All 6 tools have zero input parameters and minimal descriptions (20-150 chars). No input schemas are visible in the source code, tools accept empty objects {}. Descriptions lack WHEN/WHY context and actionable guidance. No output schemas documented. Tools are read-only reference retrievers with no error handling or recovery guidance. Naming is acceptable (verb_noun pattern: get_*) but descriptions do not meet the 50-200 char LLM-optimized baseline. No parameter descriptions exist because there are no parameters.
Get the Open Strategy Partners (OSP) editing codes documentation and usage protocol for editing texts. These semantic editing marks provide a standardized framework for content review with a teaching/learning focus.
Get the Open Strategy Partners (OSP) Web Content Meta Information Generation System for creating optimized article titles, meta titles, meta descriptions, and slugs for web content with proper keyword placement and search intent analysis.
Get the Open Strategy Partners (OSP) On-Page SEO Optimization Guide for comprehensive web content optimization, covering meta content, keyword research, content depth, search intent alignment, internal linking, and structured data.
Get the Open Strategy Partners (OSP) Product Communications Value Map Generation System for Product Positioning, including taglines, position statements, personas, value cases, and feature categorization in a structured hierarchy.
Get the Open Strategy Partners (OSP) writing guide and usage protocol for creating high-quality technical content. This guide provides systematic principles for narrative structure, flow, style, and technical accuracy.
No input schemas visible in source code. All tools accept empty input {} with no parameter definitions, types, or constraints.
Descriptions are too short (20-150 chars) and lack WHEN/WHY context. Baseline for A+ tools is 50-200 chars with explicit guidance on when to call each tool and what it returns. Current descriptions are generic reference retrievals without actionable LLM guidance.
No output schemas documented. LLMs cannot plan downstream tool calls or extract structured data without knowing what fields to expect. Tools return text content but structure is undocumented.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
Check if the server is running and can access its resources
No error handling or recovery guidance. If a markdown file is missing or unreadable, resources return generic error text. LLMs receive no actionable next steps (retry, fallback, alternative).
Tool composition is weak. All tools are read-only reference retrievers with no parameters to customize behavior. Prompts exist but are disconnected from tools, no tool accepts topic, audience, or content parameters to filter/tailor responses.