influship-api-mcp@0.22.0
The official MCP Server for the Influship API
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThis is a dangerous, default-enabled remote code-execution capability with access to environment values and a package-aligned network destination. It is not malware by itself because inspection found no lifecycle hook, hidden payload, persistence, or automatic exfiltration.
Decision evidence
public snapshot- The MCP server enables its code-execution tool by default.
- The tool runs caller-supplied TypeScript through a dynamic data-URL import.
- The subprocess receives the host environment and is granted access to environment variables and the configured API host.
- The optional HTTP transport exposes the MCP handler without an authentication check.
- package.json has no preinstall, install, or postinstall hook.
- No package code writes files, creates persistence, or performs automatic install-time execution.
- The worker limits network permission to the configured client API hostname.
Source & flagged code
5 flagged · loading sourcePackage source references dynamic require/import behavior.
instructions.jsView on unpkg · L7Source sends credentials or rich application records to a package-controlled external receiver enabled by default.
code-tool.jsView on unpkg · L137Findings
Affected versions and remediation
This report applies to influship-api-mcp@0.22.0.
See all versions of influship-api-mcp for other recorded verdicts.
- Review the evidence and your use of influship-api-mcp@0.22.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.