Conversational Ai Assistant

• Outcome: Pass
• Latest audit: May 28, 2026 at 5:49 AM UTC
• Version:
  0.1.0

Overview

This is a disclosed accounting chat-routing skill with sensitive capabilities, but its artifacts consistently require permission checks, user review, and confirmation before writes, sends, or government submissions.

Before installing, ensure the underlying OpenClaw accounting skills enforce their own permissions and confirmation gates. Treat batch actions, client communications, banking views, and government submissions as sensitive: review the affected clients, period, records, and exact action before approving anything.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The eval prompt 'Run the monthly process for all clients that are ready' is broad and maps to a potentially high-impact batch action across many tenants. In a conversational routing skill for accounting operations, overly general activation language increases the chance that ordinary user phrasing triggers sensitive workflow orchestration without sufficiently narrow scoping or explicit disambiguation.

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The intent-routing phrases are broad enough to match ordinary conversational words like "do," "run," or "start," which can cause the assistant to misclassify benign discussion as an actionable workflow request. In this skill's context, that is dangerous because the assistant is explicitly designed to orchestrate other skills and trigger operational accounting workflows, including preparation steps that touch sensitive compliance processes.

VirusTotal

65/65 vendors flagged this skill as clean.