n8n-nodes-sunlink

AI Security Review

Evidence context: threat category none; evidence quality moderate.

Security Analysis: n8n-nodes-sunlink

Tool Poisoning Assessment

No tool-poisoning findings detected. The findings summary explicitly shows "tool-poisoning":"0". This is the defining threat for MCP servers and AI agent packages, and its absence is a strong indicator of benign behavior. Unlike malicious packages that embed hidden AI directives in tool descriptions, this package contains no evidence of manipulation instructions, XML-style tags, or Unicode steganography aimed at AI agents.

Credential and Network Access

No credential theft or exfiltration architecture detected. The findings summary shows "secret":"0" with zero credential-access findings. There are no detections targeting sensitive paths like .ssh/, .aws/credentials, .kube/config, or application_default_credentials.json. Additionally, the findings summary shows "network":"0" — no NET-FETCH or NET-SOCKET_IO findings exist. The combination of credential reads plus network calls to unknown domains (the exfiltration signal) is absent.

IoC Findings Analysis

All 23 IoC findings are false positives from the XIOC extractor:

• XIOC-DOMAIN-date.now and XIOC-DOMAIN-this.name are JavaScript property access patterns misidentified as domains
• XIOC-URL-http://www.w3.org/2000/svg and XIOC-URL-http://www.w3.org/1999/xlink are standard W3C XML namespaces
• XIOC-DOMAIN-sunlinkapi.credentials.js.map is a source map filename, not a domain
• [email protected] is a placeholder email address
• XIOC-DOMAIN-docs.n8n.io and XIOC-URL-https://docs.n8n.io/credentials/' reference official n8n documentation
• XIOC-DOMAIN-github.com and XIOC-URL-https://github.com/DuanPeng1314/SunLink-n8n.git reference the package's GitHub repository
• XIOC-URL-https://www.dp7575.com.cn/api' is the only potentially real domain, but it aligns with the package's stated purpose ("SunLink (Anheyu Blog) API") and appears in configuration, not exfiltration code

Code-Smell Findings

The 19 code-smell findings ("code-smell":"19") are YARA rule noise from generic JavaScript patterns. These are classified as severity:low and do not indicate malicious behavior. No malware-signature or malware findings exist ("malware-signature":"0", "malware":"0").

Strongest Counterargument

The package has zero users ("user_count":0) and references a Chinese domain (dp7575.com.cn). However, zero users is common for new or niche n8n nodes, and the domain matches the package's documented purpose (SunLink API integration). Without network findings showing actual data transmission to this domain, or credential-access findings showing secret harvesting, there is no evidence of malicious intent. The package appears to be a legitimate workflow automation node for the SunLink blog platform.

Conclusion

This package exhibits no indicators of malicious behavior. All findings are explainable as false positives from IoC extraction artifacts and code-smell rules. The absence of tool-poisoning, credential theft, malware signatures, and network exfiltration findings supports a false positive verdict.

Key Reasons

• Zero tool-poisoning findings (the defining MCP threat)
• Zero credential-access or secret findings
• Zero network findings (no exfiltration architecture)
• All IoC findings are false positives from XIOC extractor noise
• Zero malware-signature or malware findings

False Positive Considerations

• XIOC-DOMAIN-date.now and this.name are JavaScript patterns misidentified as domains
• W3C namespace URLs (w3.org/2000/svg, w3.org/1999/xlink) flagged as IoCs
• Source map filename (sunlinkapi.credentials.js.map) misidentified as domain
• Code-smell findings from generic YARA rules on JavaScript patterns

Reviewed 2026-04-27; recommended action: suppress false positive; model confidence 72%.