@mseep/mcp-multiply-server
The AI review rates the findings as likely false positive, but the risk score (64/100) still counts them.
Analysis record
- Analysed
- 7 months ago 23 April 2026
- Version
- v1.0.0
- Artifact
- SHA256 5af80ca4…ba5793
- Detail
- Full list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
YARA Rule Matches
7 rules| Severity | Rule | Hits | Files | Metadata |
|---|---|---|---|---|
| HIGH | postinstall network communication Network communication detected | 4 | build/index.jssrc/index.tsCLAUDE.md +1 more | Risky Plugins Authors FP 30% |
| HIGH | postinstall file manipulation File system manipulation detected | 2 | build/index.jssrc/index.ts | Risky Plugins Authors FP 20% |
| HIGH | postinstall environment access Environment variable access detected | 2 | build/index.jssrc/index.ts | Risky Plugins Authors FP 40% |
| HIGH | postinstall system command System command execution detected | 4 | build/index.jssrc/index.tsCLAUDE.md +1 more | Risky Plugins Authors FP 10% |
| HIGH | postinstall crypto operations Cryptographic operations detected | 13 | build/index.jsCLAUDE.mdREADME.md +10 more | Risky Plugins Authors FP 30% |
| HIGH | postinstall obfuscation Code obfuscation techniques detected | 9 | src/sjcl.jsbuild/tweetnaclCrypto.jstest/crypto.test.ts +6 more | Risky Plugins Authors FP 20% |
| HIGH | credential env files Environment configuration file path detected | 2 | build/index.jssrc/index.ts | Risky Plugins Authors FP 10% |
Finding Categories
YARA Rules Matched
7 rules(36 hits)MCP Server Analysis
MCP servers expose tools and resources to AI assistants. Unlike browser extensions, they run as standalone processes with direct system access. Tool definitions are analyzed for prompt injection, data exfiltration, and tool poisoning patterns.
AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality strong.
Tool Poisoning Assessment
No tool-poisoning findings detected. The findings summary explicitly shows "tool-poisoning":"0". This is the most critical finding for MCP security, and its absence is significant. There are no hidden AI manipulation directives, no XML-style instruction tags, and no invisible Unicode characters encoding commands aimed at AI agents. The package description states it provides a "multiplication tool" - a simple, benign functionality that does not require hidden instructions.
Credential and Network Access
Zero credential/secret findings. The findings summary shows "secret":"0" and no credential-access findings targeting sensitive paths like .ssh/, .aws/credentials, .kube/config, or specific secret names (GITHUB_TOKEN, AWS_SECRET_ACCESS_KEY). There is no evidence of credential harvesting.
Network findings are benign. The 2 network findings and 17 IoC findings consist of:
- Legitimate URLs:
https://github.com/modelcontextprotocolandhttps://github.com/anton10xr/gibber-mcpreference the MCP protocol specification and a related open-source project - npm documentation:
https://www.npmjs.com/package/sjclreferences the SJCL crypto library - Property access false positives:
req.params.id,sjcl.bn,nacl.secretbox.open,limbs.map,decipher.final,date.noware JavaScript property chains misidentified as domains - IP
104.248.174.57with SSE endpoint could be legitimate infrastructure
There is no combination of credential reads plus network calls to unknown domains that would indicate exfiltration architecture.
Malware Signature Findings
The 36 malware-signature findings are almost certainly from bundled dependencies. MCP packages typically bundle node_modules into dist/ files, and each library triggers its own YARA matches. The absence of code-smell findings ("code-smell":"0") and obfuscation findings ("obfuscation":"0") further indicates these are not active threats but library code being scanned.
Strongest Counterargument
The developer "skydeckai" is unknown, the user count is 0, and the version is "unknown". These are weak signals of a potentially unvetted package. However, unknown publishers alone do not constitute malicious behavior. The actual code analysis shows no exfiltration, no tool poisoning, no credential theft, and no suspicious network activity. The high finding count (67 total) is driven by malware-signature matches on bundled code, not by active malicious behavior.
Conclusion
This package is a simple multiplication MCP server with findings driven by scanner noise on bundled dependencies. The zero tool-poisoning count and zero credential findings are definitive indicators of benign behavior.
Key Reasons
- Zero tool-poisoning findings - no hidden AI directives detected
- Zero credential/secret findings - no sensitive path access
- IoC findings are property access false positives and legitimate URLs
- 36 malware-signature findings from bundled dependencies, not active threats
- No exfiltration architecture (credential reads + unknown domain network calls)
False Positive Considerations
- Malware-signature findings from bundled node_modules in dist/ files
- IoC property access chains misidentified as domains (sjcl.bn, nacl.secretbox.open, req.params.id)
- Legitimate GitHub and npm URLs flagged as IoCs
- High finding count from library code, not malicious behavior
Reviewed 2026-04-27; 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