Loading ToolBench

Reading the current corpus…

iconserve MCP server · C · ToolBench

iconserve

Agent-readable static icon service (Lucide, Heroicons, Tabler, Simple Icons) on Cloudflare Workers. Provides 10,000+ open-source icons searchable by name or natural language, with SVG/PNG rendering and MCP server endpoint.

<=2025-11-25STREAMABLE_HTTP3 tools

69/100

Definition Quality66

Protocol Readiness75

SupportabilityNot measured-

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: stateful session

Deprecated protocol patterns detected

Summary

IconServe defines 3 tools with complete input schemas and descriptions. All tools follow verb_noun naming (search_icons, get_icon, list_sets). Schemas include proper types, descriptions, and constraints (enums for format/set, integer bounds for size/stroke). Descriptions are concise (40-80 chars) and action-oriented. However, output schemas are not documented, LLMs cannot predict response structure for chaining. Error handling is absent from tool definitions. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations. Parameter descriptions lack dependency hints (e.g., 'set' parameter doesn't explain how to discover valid sets).

Tools (3)

get_iconread only50/100

Get a specific icon with optional styling (color, size, stroke, format)

list_setsread only50/100

List all available icon sets with counts, licenses, and available styles

search_iconsread only50/100

Search icons by keyword or natural language query

Issues (5)

HIGHtool ↗

Output schemas not documented. LLMs cannot predict response structure (fields, types, pagination) for downstream tool chaining or result extraction.

search_iconsget_iconlist_sets

MEDIUMtool-annotations ↗

No tool annotations despite all tools being read-only. Missing readOnlyHint and idempotentHint prevent LLMs from optimizing retry logic and caching strategies.

search_iconsget_iconlist_sets

MEDIUMtool-description ↗

Parameter 'set' lacks discovery guidance. Description does not explain how to learn valid icon sets (lucide, heroicons, tabler, simple-icons), should reference list_sets or provide enum.

search_iconsget_icon

MEDIUMrecovery-guide ↗

No error handling guidance in tool definitions. LLMs cannot recover from 'icon not found', 'invalid color format', or 'size out of range' errors.

search_iconsget_icon

Recommendations

Document output schemas for all tools. Example for search_icons: 'Returns {icons: [{name, set, description, url}], total, hasMore}' so LLMs know what fields to extract.
Add readOnlyHint: true to all tool definitions in MCP registration to signal these are safe, idempotent, and cacheable.
Enhance 'set' parameter description: 'Icon set name (lucide, heroicons, tabler, simple-icons). Call list_sets to discover available sets and their licenses.'
Add error recovery guidance to get_icon: 'If icon not found, try search_icons with a partial name. If color format invalid, use hex (#RRGGBB) or CSS color names.'
Specify 'size' parameter bounds in description: 'Pixel width/height (1 - 2048, default auto). Larger sizes may increase rendering time.'

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained

Score history

Overall score trend

First recorded score · v2 rubric

69/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	C	69	<=2025-11-25	v2