Loading ToolBench

Reading the current corpus…

claude-dotfiles MCP server · F · ToolBench

claude-dotfiles

A toolkit and configuration management system for Claude Code, including MCP server integrations, plugin bootstrapping, skill management, and dotfile setup automation.

2026-07-28+STDIO2 tools

37/100

Definition Quality29

Protocol Readiness50

Supportability42

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence

Current-spec patterns detected

Summary

Both tools lack formal input schemas and have minimal descriptions. extract-tabs has a basic description (67 chars) but no visible schema validation; responsive-audit-probe's description is vague (95 chars) and does not explain when to use it or what it returns. Neither tool documents output structure, error handling, or parameter constraints. Tool names are action-oriented (extract-, responsive-audit-) but descriptions fail to guide LLM selection or explain dependencies. No evidence of pagination, structured output, or recovery guidance.

Tools (2)

extract-tabsread only38/100

Extract all open tabs from a browser into a structured TSV format with window markers and tab metadata (title, URL)

responsive-audit-proberead only33/100

Analyze page layout for responsive design failures: document overflow, interactive elements bleeding past viewport, and unannounced scrollers

Issues (5)

CRITICALconstrained-input ↗

No input schemas visible for either tool. extract-tabs accepts 'browser' param but no type/enum/description in schema. responsive-audit-probe accepts 'page' param with no validation or format guidance.

extract-tabsresponsive-audit-probe

CRITICALtool ↗

Output schemas not documented. extract-tabs claims to return 'structured TSV format' but no field list, data types, or example structure provided. responsive-audit-probe returns analysis results but structure is undefined.

extract-tabsresponsive-audit-probe

HIGHtool-description ↗

Descriptions lack actionable context. extract-tabs does not explain when to call it vs other context-gathering tools. responsive-audit-probe does not state what 'responsive design failures' means or when an LLM should invoke it.

extract-tabsresponsive-audit-probe

HIGHtool-description ↗

Recommendations

Add formal JSON Schema for both tools: extract-tabs should define browser as enum (safari|chrome|arc|brave|edge) with description 'Browser type. If omitted, detects frontmost supported browser.' responsive-audit-probe should define page as string with format 'uri' and description 'Target page URL (https://...) or DOM selector (#id, .class).'
Document output schemas explicitly. extract-tabs: return object with fields {windows: [{id, tabs: [{title, url, active}]}], format: 'tsv'}. responsive-audit-probe: return {issues: [{type, element, description, severity}], passed: boolean, timestamp}.
Expand tool descriptions to 100-150 chars with WHEN guidance. extract-tabs: 'Extract all open browser tabs into structured format. Use this to capture current context before starting a task or to audit open research.' responsive-audit-probe: 'Analyze a webpage for responsive design failures: overflow, interactive elements outside viewport, unannounced scrollers. Call before recommending layout changes.'
Add parameter descriptions for browser and page. browser: 'One of: safari, chrome, arc, brave, edge. Omit to auto-detect the frontmost supported browser.' page: 'Full URL (https://example.com) or DOM selector (#main, .container). Required.'
Implement error handling with recovery hints. E.g., if browser not found: 'Browser not detected. Supported: Safari, Chrome, Arc, Brave, Edge. Ensure one is running and frontmost.' If page unreachable: 'Page failed to load. Verify URL is valid and accessible, then retry.'

Score history

Overall score trend

First recorded score · v2 rubric

37/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	37	2026-07-28+	v2