ClawHub
Back to skill

Security audit

Bizauto Flow Free

Security checks across malware telemetry and agentic risk

Overview

This skill is a business automation advisor, but it can guide agents toward persistent scheduled jobs, external integrations, and payment-related automation without enough guardrails.

Install only if you want an agent to help design and potentially implement business automations. Before using it, require explicit confirmation before any cron creation, API call, email read/send, external message, database write, or payment-related step; use dry-run first and keep human approval for financial, HR, account, or customer-impacting actions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (8)

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
The skill extends beyond audit/design into broad external integrations, API calls, email access, messaging, storage, and scheduling. For a business-process advisory skill, this materially increases the attack surface and could lead the agent to access or transmit sensitive business data to external systems without sufficiently narrow scope or explicit per-action approval.

Intent-Code Divergence

Medium
Confidence
84% confidence
Finding
The document claims the free version is focused on audit and design, yet earlier sections describe concrete implementation and execution such as cron creation and operational workflows. This mismatch can mislead users and policy systems about what the skill may actually do, increasing the chance of unintended execution.

Intent-Code Divergence

Medium
Confidence
78% confidence
Finding
The limitations section says advanced integration capabilities are restricted, but the workflow patterns still normalize related capabilities and operational behaviors. That inconsistency weakens user understanding of boundaries and may cause agents or users to treat restricted behaviors as available and acceptable.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The activation phrases are broad, generic business requests that overlap with ordinary user prompts. This raises the likelihood of unintended invocation, causing the skill to steer conversations into automation planning or execution contexts the user did not explicitly request.

Vague Triggers

High
Confidence
95% confidence
Finding
The declared coverage list is extremely broad and vague, spanning many common business and LLM usage terms without precise trigger boundaries. This creates a high risk of accidental activation and privilege overreach, especially because the skill has exec access and discusses operational actions.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill provides creation of executable cron tasks and automated business actions without an upfront warning that these may change systems, process records, or schedule recurring operations. Users may treat the output as harmless guidance when it can directly lead to persistent or state-changing automation.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The integration section discusses API access, messaging, email, and credential handling, but lacks a direct user-facing warning that business data may be transmitted to third-party systems. This omission can lead to accidental disclosure of sensitive operational, financial, or HR data during integration workflows.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
action: "decide"
      rules:
        - condition: "amount < 5000"
          goto: "step_auto_approve"
        - condition: "amount >= 5000"
          goto: "step_manager_review"
Confidence
86% confidence
Finding
auto_approve

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.