AI Agent Security Scanner

• Outcome: Review
• Latest audit: Jul 16, 2026 · 6:50 AM UTC
• Version: 1.0.0

Overview

This is a security-audit skill that mostly matches its purpose, but its default AI analysis can send detected findings, including possible secrets or code context, to DeepSeek without a clear just-in-time warning.

Install only if you are comfortable with the tool reading the target files you scan and, when AI analysis is enabled, sending detected findings to the configured DeepSeek-compatible API. Use --no-ai for local-only scans, review generated reports for secrets, and avoid scanning proprietary or regulated code with remote AI enabled unless your organization permits that data flow.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 93% confidence
• Finding: The skill declares operational requirements for environment access and presents commands that read directories, write report files, and invoke a Python script, but it does not declare corresponding permissions. This creates a transparency and least-privilege problem: users and platforms may underestimate the skill's actual capabilities, increasing the risk of unintended file access, report overwrites, or shell-driven misuse if the implementation follows the documented behavior.

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The optional AI analysis sends discovered findings, which may include code snippets and potentially sensitive material, to an external LLM API. In a security-audit tool, scanned content is especially likely to contain secrets or proprietary code, so transmitting it off-host increases confidentiality and compliance risk.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The tool transmits scan findings to an external API without a prominent user-facing warning at the time of use. Because the findings may include extracted secrets and code context, users may unknowingly exfiltrate sensitive data during a local security scan.

External Transmission

Medium

• Category: Data Exfiltration

Content

return None
    
    if not base_url:
        base_url = os.environ.get('DEEPSEEK_BASE_URL', 'https://api.deepseek.com/v1')
    
    client = OpenAI(api_key=api_key, base_url=base_url)

• Confidence: 88% confidence
• Finding: https://api.deepseek.com/

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.