Loading ToolBench

Reading the current corpus…

declaude MCP server · C · ToolBench

← Browse

declaude

github.com/tenkenco/declaude

Claude-English to plain English translation service (gateway API + MCP)

2026-07-28+STREAMABLE_HTTP2 tools

64/100

Definition Quality67

Protocol Readiness72

Supportability54

Spec posture

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

Static source inference · medium confidence · detected: Dynamic Client Registration

Deprecated protocol patterns detected

Summary

Two tools with clear, concise descriptions and proper JSON schemas. Both tools have well-formed input schemas with type definitions. However, parameter descriptions are minimal (translate tool has only 'The text to de-Claudify.' for its single param; usage tool has no parameters). No output schemas documented. No error handling guidance. Tool names are action-oriented (translate, usage) but lack verb prefixes common in production tools (get_usage, translate_text). Schemas are present and valid but sparse.

Tools (2)

translateread onlyauthsource verified70/100

Rewrite Claude-English (AI-assistant writing tics) into plain, natural English while preserving meaning exactly.

usageread onlyauthsource verified63/100

Report this account's declaude quota for the current month: plan, translations and documents used against their limits, and an upgrade link when on the free tier.

Issues (5)

HIGHtool ↗

No output schemas documented. LLMs cannot plan downstream operations or extract structured data from responses.

translateusage

HIGHtool-description ↗

Parameter descriptions are minimal. 'The text to de-Claudify' is too terse; should explain format, length limits, and expected content type.

translate

MEDIUMtool ↗

Tool names lack action verbs. 'translate' and 'usage' are weak; should be 'translate_text' and 'get_usage' to match production conventions.

translateusage

MEDIUMrecovery-guide ↗

No error handling guidance. If translate fails (e.g., text too long, invalid encoding), LLM has no recovery path.

translate

MEDIUMconstrained-input ↗

No input validation constraints documented. 'text' parameter has no length limit, format, or encoding specification in description.

Recommendations

Add output schema documentation for both tools. E.g., translate should document: {translated_text: string, confidence: number, changes_made: string[]}.
Expand translate parameter description: 'The text to de-Claudify (max 10,000 chars, UTF-8 encoded). Preserves meaning while removing AI-assistant writing patterns.'
Rename tools to verb_noun convention: 'translate_text' and 'get_usage' for clarity and LLM discoverability.
Add error handling guidance in descriptions: 'If text exceeds 10,000 chars, returns 413 with suggestion to split into paragraphs.'
Document usage response structure: plan tier, monthly limit, current usage count, remaining quota, and upgrade URL for free tier.

Score history

Overall score trend

First recorded score · v2 rubric

64/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	C	64	2026-07-28+	v2