Loading ToolBench

Reading the current corpus…

mcp-project MCP server · F · ToolBench

mcp-project

A collection of MCP servers including text analysis, Fibonacci calculations, distance unit conversion, and academic paper research tools powered by arXiv integration

2026-07-28+STDIO6 tools

43/100

Definition Quality58

Protocol Readiness30

Supportability26

Spec posture

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

Static source inference · medium confidence

Current-spec patterns detected

Summary

This MCP server exhibits significant gaps in definition quality. While all 6 tools have basic descriptions and schemas, the descriptions are minimal (10-50 chars, well below the 194-char baseline), parameters lack depth, and output schemas are undocumented. Tool naming is acceptable but descriptions do not follow the LLM-optimized pattern of stating WHAT, WHEN, and WHY. The server shows minimal error handling guidance and no recovery patterns. Most tools score in the 35-50 range, pulling the overall average down.

Tools (6)

analyze_textread onlysource verified60/100

Perform comprehensive text analysis.

convert_distanceread onlysource verified61/100

Convert between different distance units.

extract_inforead onlysource verified51/100

Search for information about a specific paper across all topic directories.

fibonacci_sumread onlysource verified53/100

Calculate Fibonacci sequence and its sum up to n numbers.

get_weatherread onlysource verified59/100

Returns the current weather condition and temperature in Celsius for a given city.

search_paperswritesource verified61/100

Search for papers on arXiv based on a topic and store their information.

Issues (8)

CRITICALD. SCHEMAS & OUTPUT ↗

Output schemas completely undocumented. LLMs cannot infer what fields exist in tool responses, forcing guesswork on downstream parameter passing.

analyze_textfibonacci_sumconvert_distancesearch_papersextract_infoget_weather

CRITICALB. DESCRIPTIONS ↗

Tool descriptions are 25 - 50 characters, far below the 194-char baseline. They lack WHEN to use the tool, WHAT it modifies, and WHY an LLM should select it.

fibonacci_sumextract_infoanalyze_textget_weathersearch_papersconvert_distance

HIGHC. PARAMETERS ↗

No enums for constrained parameters. convert_distance lists units as freeform string descriptions; no schema enum prevents hallucinated unit names like 'kilometer' or 'km/s'.

convert_distance

HIGH

Recommendations

Expand ALL tool descriptions to 50 - 200 characters. Format as: 'Performs [ACTION] on [RESOURCE] based on [INPUTS]. Use this when [USE CASE]. Returns [OUTPUT STRUCTURE]. Note: [SPECIAL BEHAVIOR, e.g., stateful writes].' Example: 'Searches arXiv for papers matching a topic and saves results to a local JSON file. Use this to discover recent research; results are persisted for later retrieval. Returns list of paper IDs found. Note: This tool writes to disk.'
Document output schemas in code docstrings. Add 'Returns: Dict[str, Union[int, float]] with keys: word_count (int), char_count (int), sentence_count (int), avg_word_length (float), avg_sentence_length (float), unique_words (int).' to each tool.
Add enum constraints to convert_distance: from_unit and to_unit should be declared as enums in JSON Schema ('enum': ['km', 'mi', 'm', 'cm', 'ft']) rather than relying on freeform string descriptions.
Add input validation and bounds. fibonacci_sum should reject n < 1 with message: 'n must be >= 1. Got: {n}.' convert_distance should validate value > 0.
Enhance extract_info description: 'Searches all stored paper directories for a specific paper ID (arXiv short ID format, e.g., 2401.12345) and returns its metadata (title, authors, summary, PDF URL, published date). Use this after search_papers to retrieve full details. Returns JSON string if found, error message if not.'
Add error recovery hints. Replace 'Invalid unit. Use: m, km, mi, ft' with 'Invalid unit: {invalid_unit}. Must be one of: m, km, mi, ft. Example: convert_distance(value=100, from_unit="m", to_unit="km")'.

Score history

Overall score trend

↓ 6 points across a rubric change (v1 → v2)

43/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-22	F	43	2026-07-28+	v2
2026-03-09	F	49	-	v1