MCP Proxy
The AI review rates the findings as likely false positive, but the risk score (41/100) still counts them.
Analysis record
- Analysed
- 10 months ago 12 January 2026
- Version
- v0.0.6
- Artifact
- SHA256 354f0b5f…fed869
- Detail
- Full list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
YARA Rule Matches
8 rules| Severity | Rule | Hits | Files | Metadata |
|---|---|---|---|---|
| HIGH | postinstall crypto operations Cryptographic operations detected | 1 | dist/extension.js | Risky Plugins Authors FP 30% |
| HIGH | postinstall network communication Network communication detected | 4 | dist/extension.jsREADME.mdCHANGELOG.md +1 more | Risky Plugins Authors FP 30% |
| HIGH | postinstall file manipulation File system manipulation detected | 1 | dist/extension.js | Risky Plugins Authors FP 20% |
| HIGH | postinstall persistence mechanism Persistence mechanism detected | 1 | package.json | Risky Plugins Authors FP 20% |
| HIGH | postinstall system command System command execution detected | 3 | dist/extension.js.github/copilot-instructions.mdREADME.md | Risky Plugins Authors FP 10% |
| HIGH | postinstall file download File download activity detected | 1 | package.json | Risky Plugins Authors FP 30% |
| HIGH | postinstall obfuscation Code obfuscation techniques detected | 1 | dist/extension.js | Risky Plugins Authors FP 20% |
| HIGH | credential env files Environment configuration file path detected | 1 | dist/extension.js | Risky Plugins Authors FP 10% |
Publisher Evidence
Limited evidenceViktor Borg
Publisher identity, store signals, distribution reach, and warning signals used for context. Treat this as supporting evidence, not a clean bill of health.
13 evidence rows available.
Finding Categories
YARA Rules Matched
8 rules(13 hits)AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality moderate.
The MCP Proxy extension by borgius presents 64 total findings, but all originate from known false positive sources. The 49 IoC findings are exclusively XIOC property access chain misclassifications. Findings like XIOC-DOMAIN-t.id, XIOC-DOMAIN-c.name, XIOC-DOMAIN-s.stream, XIOC-DOMAIN-t.data, and XIOC-DOMAIN-c.inputschema.properties are JavaScript property access patterns being misread as network domains. The CVEQ documentation explicitly lists this as a garbage pattern: "Property access chains misread as domains: b.call, h.next, g.id". The one legitimate URL, XIOC-URL-http://keepachangelog.com/, references a standard changelog documentation site, not malicious infrastructure.
The 13 code-smell findings represent standard YARA noise on JavaScript code. Per CVEQ guidelines, "code-smell findings (severity=low, finding_type=code-smell) are NOISE" and should never drive a verdict. Zero malware signatures, zero secret/credential findings, and zero network findings indicate no actual malicious behavior.
The extension's stated purpose—proxying MCP (Model Context Protocol) servers as VS Code Language Model Tools—is legitimate development functionality. MCP is an emerging standard for connecting AI tools to development environments, and proxying these servers requires normal IDE extension capabilities. No manifest-analysis findings exist in the evidence, so actual workspace access permissions cannot be assessed from this data.
Strongest counterargument to this verdict: The low user count (30) and recent version (0.0.6) could indicate a new supply chain attack attempting to establish presence before detection. However, supply chain attacks require evidence of malicious capability—credential theft, exfiltration, or postinstall payload execution. This extension has zero secret findings, zero network findings, and zero malware signatures. The findings are clearly from documented false positive sources, not actual threats. Without evidence of malicious intent or capability, the low user count alone does not justify a risk classification.
Recommended action is to suppress these false positives and document the XIOC property access chain pattern for future triage.
Key Reasons
- All IoC findings are XIOC property access chain false positives
- Zero malware signatures detected
- Zero credential/secret findings
- Code-smell findings are documented noise
- Extension purpose is legitimate MCP proxy functionality
False Positive Considerations
- XIOC property access chain false positives (t.id, c.name, s.stream)
- Code-smell YARA noise on JavaScript
- Bundled dependency findings
- No actual malware signatures or credential theft patterns
Reviewed 2026-04-28 · version 0.0.6; recommended action: suppress false positive; model confidence 85%.
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