ClawHub
Back to skill

Security audit

Ba Workbench International

Security checks across malware telemetry and agentic risk

Overview

This is a text-only business-analysis helper with broad templates and advice, and no evidence of hidden execution, data access, persistence, or exfiltration.

Install this if you want a large, opinionated business-analysis template library. Be aware that it may answer broadly framed BA requests and append lengthy author/license disclaimers to outputs. The package publisher should fix the ba-workbench versus ba-workbench-international metadata mismatch, but no malicious behavior was found.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The metadata declares the installed skill as "ba-workbench" while the provided skill context identifies it as "ba-workbench-international", creating an identity mismatch. This can confuse provenance, dependency tracking, policy enforcement, or review workflows, and in some systems may allow a package to masquerade as a different trusted skill or bypass name-based controls.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The README says the skill can be triggered by broadly phrased natural-language requests like 'Run an industry analysis' or 'Write a Business Case.' Overly broad invocation criteria can cause accidental activation in unrelated contexts, increasing the chance that the skill takes over a conversation or injects business-analysis behavior when the user did not explicitly intend to invoke it.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The skill advertises itself as an all-purpose 'external brain' and invites users to 'tell me directly what business analysis you need,' which creates very broad activation criteria. In agent ecosystems, vague invocation boundaries can cause the skill to be selected for routine or adjacent requests where its heavy prescriptive behavior, legal boilerplate, and embedded workflow assumptions are not appropriate.

Natural-Language Policy Violations

Medium
Confidence
79% confidence
Finding
The front matter fixes the skill language to en-US without offering user choice. In isolation this is not severe, but in a multi-locale agent environment it can override user preference, cause misalignment with regional compliance or terminology needs, and reduce safe, accurate communication for non-English users.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.