Based on the RiskyPlugins AI security review of the observed evidence.
No individual score drivers were recorded for this analysis.
Analysis record
- Analysed
- 4 months ago 1 July 2026
- Version
- v10.0.0
- Artifact
- SHA256 d8de197a…a22bd9
- Detail
- Full list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
Finding Categories
AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality strong.
Tool-Poisoning Assessment: Zero tool-poisoning findings were detected in this package. The scanner found no hidden AI directives, no XML-style instruction tags, and no tool descriptions containing manipulation instructions. This is a standard database library, not an MCP tool framework, so the absence of tool-poisoning findings is expected and appropriate.
Credential and Network Access: The evidence shows zero secret findings and zero network findings. There is no code reading .ssh, .aws, .kube, or other sensitive credential paths. There is no HTTP POST/fetch to unknown domains. The package reads standard configuration for a database library but does not exfiltrate data.
IoC Findings Analysis: All 226 IoC findings are documentation URLs incorrectly extracted from markdown files. Specific examples include XIOC-URL-https://github.com/Level/webpack-starter, XIOC-URL-https://nodejs.org/api/n-api.html, and XIOC-URL-https://img.shields.io/node/v/level.svg. These are legitimate GitHub repository links, Node.js API documentation, and npm badge URLs from the package's README and documentation. The IoC extractor treats any URL string as a finding, which is a known false-positive pattern in the CVEQ scanner.
Code-Smell Findings: Eight low-severity code-smell findings exist, which is expected for any non-trivial JavaScript package. These are classified as low severity and match standard Node.js patterns (fs, process.env, crypto) that the YARA rules flag indiscriminately.
Strongest Counterargument: The high total finding count (237) could suggest malicious behavior. However, this is entirely explained by the IoC extractor flagging every URL in documentation as a finding. None of the 226 IoC findings point to suspicious exfiltration domains—every single one is a legitimate GitHub, npm, or shields.io URL. The absence of tool-poisoning (0), secret access (0), and network activity (0) findings confirms this is a false positive from documentation parsing, not an exfiltration architecture.
Conclusion: This is the Level database, a well-known npm package for Node.js and browsers. The findings represent scanner noise from documentation URLs, not malicious behavior.
Key Reasons
- Zero tool-poisoning findings detected
- Zero secret/credential access findings
- Zero network findings to unknown domains
- All 226 IoC findings are benign documentation URLs (GitHub, npm, shields.io)
- Package is a well-known database library, not an MCP tool
False Positive Considerations
- IoC extractor flagging documentation URLs as findings
- Code-smell YARA rules matching standard Node.js patterns
- URL strings in README/markdown files incorrectly classified as IoCs
Reviewed 2026-04-27 · version 10.0.0; recommended action: suppress false positive; model confidence 92%.
Source Code Not Available
Source code is not available for this version of the extension.
About This Extension
Frequently Asked Questions
Similar Extensions
Related extensions from the same publisher or marketplace