MCP Registry

toolsdk

by seeygo
5f2d6507-6812-547e-a4e5-510c1b0ffdd5 | Store v2.1.0
100/ 100
CRITICAL risk
Risk verdict
Do not install

Score-based assessment (critical risk, 100/100). An AI review is available in the AI review tab.

Analysis record

Analysed
2 months ago 26 August 2026
Version
v2.1.0
Artifact
SHA256 b9824900…7b318b
Detail
Full list

Evidence ledger

Ranked by severity · findings with a source location link to the code viewer

227 findings
Showing 25 of 148 · highest severity first

YARA Rule Matches

8 rules
SeverityRuleHitsFilesMetadata
LOWpostinstall environment access 14
dist/types/dto-integration.d.tsdist/types/dto-component.d.tsdist/types/vo-template.d.ts +11 more
-
LOWpostinstall network communication 18
dist/api/package/configuration-instance.d.tsdist/types/dto-component.d.tsdist/types/bo-zapier.d.ts +15 more
-
LOWpostinstall file manipulation 19
dist/types/dto-package.d.tsdist/types/bo-configuration.d.tsdist/types/bo-package-data.d.ts +16 more
-
LOWpostinstall crypto operations 1
dist/react/index.esm.js
-
LOWpostinstall registry modification 2
README.mddist/types/index.esm.js
-
LOWpostinstall obfuscation 2
dist/react/index.esm.jsdist/types/index.esm.js
-
LOWpostinstall system command 5
dist/api/package/package.d.tsREADME.mddist/api/index.esm.js +2 more
-
LOWpostinstall file download 18
dist/types/bo-package-data.d.tsdist/types/bo-configuration.d.tsdist/types/dto-component.d.ts +15 more
-

Finding Categories

147
Secrets
34
IoC Indicators

YARA Rules Matched

8 rules(79 hits)
postinstall environment access postinstall network communication postinstall file manipulation postinstall crypto operations postinstall registry modification postinstall obfuscation postinstall system command postinstall file download

MCP Server Analysis

MCP servers expose tools and resources to AI assistants. Unlike browser extensions, they run as standalone processes with direct system access. Tool definitions are analyzed for prompt injection, data exfiltration, and tool poisoning patterns.

AI Security Report

AI Security Review

Summary

ToolSDK.ai MCP server scores CRITICAL with tool-poisoning detected, 79 malware signatures, and 217 IoC matches. The credential-access findings are spread across type definition files (.d.ts) which is likely pattern-matching on field names like credentials in an integration SDK — expected for a platform offering 100+ third-party integrations. The 2 tool-poisoning findings warrant direct inspection. Malware signatures and IoC volume need source review before escalating.

Key Reasons

  • Tool-poisoning detected (2 findings) — direct risk for AI agent manipulation; requires manual inspection of the specific tool descriptions
  • 79 malware signatures fired — high volume but no confirmed malware_detected flag; likely broad YARA rules hitting minified SDK bundle
  • 217 IoC matches on an integration SDK are plausible (100+ external service endpoints) but need filtering for genuinely malicious indicators
  • Credential-access findings are all in .d.ts type definitions referencing a credentials field — almost certainly a false positive from keyword matching on a legitimate auth abstraction

Reviewed 2026-03-13 · version 2.1.0; recommended action: manual reanalysis; model confidence 72%.

Source Code Viewer

Source Code Not Available

Source code is not available for this version of the extension.

About This Extension

ToolSDK.ai - A decentralized Tool SDK for AI Agents. Add automation features and 100+ integrations to your AI apps.

Frequently Asked Questions