Regex Assistant

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:40 PM UTC
• Version:
  1.0.0

Overview

This is a disclosed AI regex tool; its main risk is that AI commands can send your regex inputs and selected file content to EvoLink.

Install only if you trust EvoLink with the regex prompts, sample strings, and file excerpts you choose to analyze. Use the local cheatsheet for offline work, and do not run AI test/debug commands on credentials, private logs, customer data, proprietary source, or other confidential files.

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 (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill declares required binaries and environment variables and its documented commands invoke shell scripts that can read local files, use environment secrets, and make outbound network requests, but it does not declare explicit permissions for those capabilities. This creates a transparency and policy-enforcement gap: users or platforms may not realize the skill can access file contents and transmit them to a third-party API, especially for the `test` command that sends file content for remote analysis.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The `test` command reads user-supplied file contents and sends up to 8000 characters to the external EvoLink AI API for analysis, but the script does not provide a clear, explicit disclosure or consent prompt at the point of transmission. This creates a real data-leak risk because users may pass sensitive files assuming processing is local, while the command silently exfiltrates contents to a third party.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The test flow reads arbitrary local file contents and includes them in the prompt sent to the external EvoLink API. While this is core functionality rather than overtly malicious behavior, there is no explicit consent gate, redaction, or strong warning at the point of use, so users may unintentionally upload sensitive data from local files to a third party.

Ssd 3

Medium

• Confidence: 98% confidence
• Finding: This command explicitly loads a user-specified file, captures up to 8000 bytes of its contents, and transmits that content to an external AI service for analysis. In a developer tool this can expose secrets, credentials, source code, customer data, or proprietary text if the user points it at a sensitive file without understanding the data-sharing behavior.

VirusTotal

61/61 vendors flagged this skill as clean.