ClawHub
Back to skill

Security audit

Fireflies

Security checks across malware telemetry and agentic risk

Overview

This Fireflies skill is a disclosed OOMOL connector wrapper with expected account access and clear cautions for write and delete actions.

Install only if you trust OOMOL to broker access to your Fireflies account. Review the Fireflies scopes you connect, confirm any write or destructive action before it runs, and be cautious with raw GraphQL requests because the local skill text does not show the backend enforcement that keeps them read-only.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Intent-Code Divergence

Medium
Confidence
85% confidence
Finding
The skill advertises `execute_graphql_query` as read-only, but the file shows no local guardrails, validation, or allowlisting that would actually restrict the supplied GraphQL document to queries only. If the backend action is not strictly enforced server-side, an agent could be induced to send mutations or other overly broad requests, leading to unintended data modification or sensitive data access.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.