@bretterplane/mcp-tools
The AI review rates the findings as likely false positive, but the risk score (100/100) still counts them.
Analysis record
- Analysed
- 2 months ago 26 August 2026
- Version
- v0.3.1
- Artifact
- SHA256 b7758aa3…a8bae1
- Detail
- Full list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
Finding Categories
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: The scanner found 0 tool-poisoning findings. This package is an SDK library for building MCP clients and servers, not an MCP server itself. The findings_summary explicitly shows "tool-poisoning":0. There is no evidence of hidden AI manipulation directives, invisible Unicode characters, or XML-style instruction tags in tool descriptions.
Credential and Network Access: The 10 secret findings are code-smell matches on API key patterns in documentation or configuration examples, not actual credential harvesting. The 7 network findings correspond to normal SDK functionality. Critically, there are no credential-access findings targeting sensitive paths like .ssh/, .aws/credentials, .kube/config, or application_default_credentials.json. The 154 IoC findings are all documentation URLs: XIOC-URL-https://modelcontextprotocol.io/specification/draft/basic/authorization links to MCP specification docs, XIOC-URL-https://github.com/modelcontextprotocol/typescript-sdk points to the official MCP SDK repository, XIOC-URL-https://github.com/clerk/mcp-tools references Clerk's MCP tools, and XIOC-URL-https://yourapp.com is a placeholder URL in examples. None are exfiltration endpoints.
Strongest Counterargument: The package has 242 total findings and 0 users, which could suggest an unmaintained or suspicious package. However, this is a development library (version 0.3.1) where high finding counts are expected from bundled code and documentation. The developer bretterplane is not well-known, but the package content is legitimate SDK code referencing official MCP and Clerk resources. The absence of tool-poisoning findings, malware signatures (0), and obfuscation (0) confirms this is a false positive driven by the scanner treating documentation URLs and library patterns as security concerns.
Conclusion: This is a benign MCP development library. All findings are false positives from documentation URLs, code-smell rules matching normal patterns, and the scanner's inability to distinguish SDK tool definitions from actual tool poisoning.
Key Reasons
- Zero tool-poisoning findings in findings_summary
- All 154 IoCs are documentation URLs, not exfiltration endpoints
- Zero malware signatures and zero obfuscation findings
- Package is an SDK library, not an executable MCP server
- No credential-access findings targeting sensitive paths (.ssh, .aws, .kube)
False Positive Considerations
- Documentation URLs extracted as IoCs (modelcontextprotocol.io, github.com, clerk.com)
- Code-smell rules firing on normal SDK library patterns
- Secret findings matching API key patterns in config examples, not actual credential access
- Zero tool-poisoning findings confirms no hidden AI directives
Reviewed 2026-04-27 · version 0.3.1; recommended action: suppress false positive; model confidence 88%.
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