ClawHub
Back to skill

Security audit

招投标智慧问答助手Bidding Qa Assistant

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Chinese bidding and tendering Q&A assistant that uses scoped IMA knowledge-base retrieval and does not show hidden installation, persistence, exfiltration, or destructive behavior.

Install only if you want a Chinese-language bidding/tendering compliance assistant that queries IMA knowledge bases. Do not upload confidential bid documents unless you are comfortable having their key contents parsed and used to form IMA search queries; verify legal conclusions against official sources or counsel for high-stakes matters.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The activation criteria are overly broad and can match ordinary bidding-related requests, causing the skill to trigger outside its intended scope. In an agent setting, ambiguous routing can misapply specialized instructions, force the wrong knowledge-base workflow, and produce authoritative-seeming but context-inappropriate guidance.

Natural-Language Policy Violations

Medium
Confidence
80% confidence
Finding
The skill is written entirely for Chinese-language operation without an explicit language-selection or fallback mechanism. This can cause users in other languages to receive misunderstood, untranslated, or partially interpreted compliance guidance, which is risky in a legal and procedural domain where precision matters.

Vague Triggers

Medium
Confidence
72% confidence
Finding
The trigger conditions are broad enough that the skill may activate for ordinary bidding-related conversations and uploaded documents, causing over-collection and processing of potentially sensitive procurement materials. In this skill's context, unintended activation is more concerning because it routes files into retrieval and analysis workflows, which can expand exposure of confidential tender or complaint data beyond what the user expected.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The activation description is broad enough to trigger on many ordinary bidding, legal, or knowledge-base related requests, which can cause the skill to engage outside narrowly intended contexts. In a compliance-sensitive domain like tendering and complaints guidance, over-broad activation increases the chance of unsolicited specialized advice, misrouting user intent, or overshadowing safer/general handlers.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.