Agent Compete Scope

• Outcome: Pass
• Latest audit: May 29, 2026, 8:16 PM
• Version:
  1.0.2

Overview

This appears to be a normal competitor-analysis skill that uses web search and external LLM APIs, with privacy and API-key handling caveats users should understand.

Install only if you are comfortable sending competitor-research inputs and resulting analysis to Tavily and the selected LLM provider. Use dedicated, limited API keys, avoid confidential strategy or customer data unless those providers are approved, and review or disable the mock npm start behavior before production use.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 82% confidence
• Finding: The skill appears to rely on network and environment capabilities, but no permissions are explicitly declared. That creates a trust and review gap: operators cannot easily see that the skill may access external resources or environment-derived data, which can lead to unintended data exposure or unvetted outbound requests. In a competitor-analysis skill, network access is plausible, but undeclared capability use still weakens security transparency and policy enforcement.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The README explicitly documents use of Tavily web search and external LLM APIs but provides no notice that user inputs, competitor names, and analysis context may be transmitted to third-party services. This can create privacy, confidentiality, and compliance risks because users may unknowingly send sensitive business strategy data outside their environment.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The function sends aggregated article titles, URLs, and full content to an external LLM provider without any visible consent, minimization, or classification checks. If the collected content includes proprietary research, private data, or licensed material, this can cause unintended third-party disclosure and compliance issues.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The comparison-matrix step transmits compiled product and competitor profile data to an external model, which may include sensitive strategic positioning and internal product assessments. Sharing this information without a visible warning or control can leak confidential business intelligence to a third party.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: This step sends the synthesized comparison matrix and strategic analysis context to an external LLM to generate whitespace opportunities and recommendations. Because the data is higher-level strategic intelligence, disclosure risk is amplified: it can expose inferred weaknesses, market opportunities, and internal positioning decisions.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: This code sends the combined system prompt and user message to Google's external Gemini API, which can expose sensitive user or business data to a third party if users are not clearly informed. The risk is contextual rather than exploit-code driven, but in an agent skill that may process competitive analysis inputs, proprietary strategy, customer data, or internal documents could be transmitted off-platform without explicit disclosure or consent.

VirusTotal

66/66 vendors flagged this skill as clean.