langchain4j-mcp-server-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Downloads and executes official Model Context Protocol (MCP) server implementations from the @modelcontextprotocol organization on NPM, such as the sqlite and everything server packages.
- [COMMAND_EXECUTION]: Bootstraps MCP servers using npm exec and docker run commands to start sidecar processes for tool execution.
- [COMMAND_EXECUTION]: Provides an implementation pattern for a database tool that executes arbitrary SQL queries provided as string arguments.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources (databases, GitHub, static resources) which may contain untrusted content that could influence agent behavior.
- Ingestion points: Resource providers and tool implementations (e.g., GitHubToolProvider) in references/examples.md and assets/mcp-server-template.java that fetch data from external environments.
- Boundary markers: The instructions in SKILL.md explicitly advise developers to filter available tools by trust level and sanitize content returned by external resources.
- Capability inventory: The skill patterns include capabilities for SQL execution, GitHub API interaction, and process invocation via Stdio.
- Sanitization: Instructions recommend sanitizing external content before downstream use, though the simplified code examples focus on implementation logic rather than complex validation.
Audit Metadata