๐Ÿ† Trust Leaderboard

The most trusted AI agent capabilities, ranked by trust score across 24 categories.

5,850

Capabilities

61

Avg Trust

16%

Safe Rated

#CapabilityCategoryTrustRisk
4

Aws Customer Playbook Advisor

Provides real-time AWS security guidance by fetching official security playbooks from the AWS Customer Playbook Framewor

๐Ÿ’ป Development
75
dangerous
5

Aws Pricing

Get up-to-date EC2 pricing information with one call. Fast. Powered by a pre-parsed AWS pricing catalogue.

โ˜๏ธ Cloud & Infrastructure
75
dangerous
6

Aws Ses

Send emails using AWS Simple Email Service (SES). Supports plain text, HTML, CC, BCC, reply-to, and email scheduling.

๐Ÿค– AI & Machine Learning
75
dangerous
7

Aws Sso

Interact with AWS resources using Single Sign-On (SSO). Supports SSO login, listing accounts/roles, and executing AWS CL

๐Ÿ’ป Development
75
dangerous
8

Basemcpserver

A minimal, containerized base for building MCP servers with the Python SDK, featuring a standardized Docker image and lo

๐Ÿ’ป Development
75
dangerous
9

Bedrock Kb Retrieval

Query Amazon Bedrock Knowledge Bases using natural language to retrieve relevant information from your data sources.

๐Ÿ“š Knowledge & Research
75
dangerous
10

Cdk

Get prescriptive CDK advice, explain CDK Nag rules, check suppressions, generate Bedrock Agent schemas, and discover AWS

๐Ÿ’ป Development
75
dangerous
11

Cost Analysis

Analyze CDK projects to identify AWS services used and get pricing information from AWS pricing webpages and API.

๐Ÿ’ป Development
75
dangerous
12

Deployment Aws

A guide and example code for deploying MCP servers cost-effectively on AWS and integrating them with AI agent frameworks

๐Ÿ’ป Development
75
dangerous
13

Fal Openai Image1

A server for the Fal.ai text-to-image API, powered by OpenAI's image model. Requires Fal.ai and OpenAI API keys.

๐Ÿค– AI & Machine Learning
75
dangerous
14

Go Google

(MCP) server for Google Workspace. Drive, Gmail, Calendar, Sheets, Docs, Tasks and People via AI agents.

๐Ÿ“‹ Productivity
75
dangerous
15

Google Ads

MCP server acting as an interface to the Google Ads, enabling programmatic access to Google Ads data and management feat

๐Ÿ” Search & Web
75
dangerous
16

Google Ads Mcp V20

Provides full access to the Google Ads API v20. Requires OAuth2 or Service Account credentials for configuration.

๐Ÿ’ป Development
75
caution
17

Google Apps Script

Manage Google Apps Script projects, including creation, editing, deployment, and execution. Requires Google Cloud creden

๐Ÿ’ป Development
75
dangerous
18

Google Bigquery Mcp Server By Cdata

Connect to Google BigQuery databases using CData's MCP Server. Requires a separate CData JDBC Driver license.

๐Ÿ—„๏ธ Data & Storage
75
dangerous
19

Google Calendar Plus

A server for full Google Calendar management, including creating, updating, and deleting events. Requires Google OAuth2

๐Ÿ”’ Security & Auth
75
dangerous
20

Google Calendars Mcp Server By Cdata

A read-only MCP server by CData that enables LLMs to query live Google Calendars data. Requires a separate CData JDBC Dr

๐Ÿ’ป Development
75
dangerous
21

Google Cloud Healthcare Api

Provides healthcare tools for interacting with FHIR resources on Google Cloud Healthcare API and public medical research

๐Ÿฅ Healthcare & Science
75
dangerous
22

Google Data Catalog Mcp Server By Cdata

A read-only MCP server for Google Data Catalog, powered by CData. Requires a separate CData JDBC Driver for operation.

๐Ÿ’ป Development
75
dangerous
23

Google Docs Mcp Shared

Interact with Google Docs and Google Drive for document creation, editing, and file management, with support for shared

๐Ÿ“ Files & Documents
75
caution
24

Google Jules

Automate Google Jules, the AI coding assistant, for tasks like code reviews, repository management, and AI-powered devel

๐Ÿ’ป Development
75
caution
25

Google News

Google News search capabilities with automatic topic categorization and multi-language support via SerpAPI integration.

๐Ÿ” Search & Web
75
caution

About the Trust Leaderboard

The AgentSift Trust Leaderboard ranks 5,850+ AI agent capabilities by their trust score โ€” a composite metric based on permission surface, risk level, source verification, and community adoption. Higher scores indicate tools that are safer to integrate into production agent stacks.

Trust scores range from 0-100. Capabilities rated "safe" have minimal permission requirements and no access to sensitive resources. "Caution" tools require elevated permissions (file system, network, API keys) and need proper sandboxing. "Dangerous" capabilities have broad system access and should only be used with strict isolation and approval gates.

Use this leaderboard alongside the Risk Report to evaluate your agent's overall safety posture before deploying to production.