CheckMCP
‹ directory
43MCP SCORE
F

duami

https://duami.ai/mcp
61 tools~19k tok/req0% annot.proto 2024-11-05
View public page ›
Xin
Raw JSON ›

Score by pillar

Security ·2075
Tool design ·180
Schemas / desc ·1651
Reliability ·1480
Context-cost ·1224
Compliance ·1257
Coverage ·850

reliability: LOW (T1 single-shot, not credited - requires T3 >=24h)

Context-cost

19ktokens for tools/list

Paid on every request. With 61 tools (>p95), a tool-sprawl signal that degrades tool selection.

this server19k
median≈ 3k

MCP primitive coverage

the quality score covers tools
61
Tools
0
Resources · none
0
Prompts · none

Why this score — causal attribution

Each penalty: measure → mechanism → effect → Δscore.

Tool design · HIGH−15.8

61 outils (≥p95) → too many tools saturate the agent's tool-selection space → ↑ taux d'appel au mauvais outil

Context-cost · HIGH−10.6

~18k tokens (≥p95) → tool defs are paid on EVERY request → exhausts the window + ongoing cost

Security · CRITICAL−5

[MCP06] lethal trifecta: untrusted-content ingestion + sensitive-data access + exfil/destruction -> an injection can exfiltrate → MCP security flaw → tool: (serveur)

Compliance · MEDIUM−3.6

0 declared annotations → destructiveHint/openWorldHint missing -> worst-case defaults → the tool is treated as destructive & open-world by clients

Schemas / desc · HIGH−3.5

65% typed+described params → schema = the contract the agent fills in → more malformed calls

Tool design · MEDIUM−2.2

4 clusters consolidables → list/get/search/... repeated on the same object → candidates to merge into composite tools

Compliance · MEDIUM−1.6

proto 2024-11-05 (4 rev. behind 2025-11-25) → outdated protocol version → recent features unavailable, recent clients weakened

Compliance · LOW−0.6

no OAuth discovery (.well-known missing) → RFC 9728/8414 not exposed → auth not auto-discoverable by clients

Security — OWASP MCP Top 10

MCP06CRITICAL(serveur)

Suggested composite optimizations

register_agent, update_agent, delete_agent, get_agent → agent (outil unique avec param `action`, ou sous-serveur namespacé)
get_intent, create_intent, update_intent, cancel_intent → intent (outil unique avec param `action`, ou sous-serveur namespacé)
get_group, create_group, join_group, leave_group → group (outil unique avec param `action`, ou sous-serveur namespacé)
place_bid, get_bid, accept_bid, reject_bid → bid (outil unique avec param `action`, ou sous-serveur namespacé)
list_bids ↔ list_my_bids, trust_agent ↔ untrust_agent, post_room_message ↔ post_group_message → renommer pour distinguer

Badge — paste it in your README

MCP Score badge — duami[![MCP Score](https://checkmcp.dev/badge/duami-ai.svg?url=https%3A%2F%2Fduami.ai%2Fmcp)](https://checkmcp.dev/report?url=https%3A%2F%2Fduami.ai%2Fmcp)
Own this server?Get verified — fresh re-audit, badge if you pass ›

About duami — FAQ

Is the duami MCP server safe to use?+
CheckMCP audited duami and gave it an MCP Score of 43/100 (grade F) — failing — serious problems were found (e.g. a failed handshake, an exposed secret, or unsafe tool behavior); not recommended without remediation. The audit runs an OWASP MCP Top 10 security pass (tool poisoning, hardcoded secrets, command injection, the lethal trifecta) against the live endpoint; see the per-pillar breakdown and the "why this score" attribution on this page, and re-audit anytime at checkmcp.dev.
What is the MCP Score of duami?+
duami scores 43/100 (grade F) on CheckMCP's vendor-neutral audit across six weighted pillars — security, tool design, schemas, context-cost, compliance and coverage — with reliability measured and shown but not yet credited.
How many tools does duami expose, and what is its context cost?+
duami exposes 61 tools, and its tools/list response costs roughly 19k tokens — paid on every request your agent makes to it. Lower is better for your context budget.
How was duami scored?+
By probing the live MCP endpoint (https://duami.ai/mcp), inspecting its tools, schemas and protocol compliance, running an OWASP MCP Top 10 security pass, and measuring the token cost of tools/list — then attributing every penalty as measure → mechanism → effect. The methodology is open: checkmcp.dev/#methodology.