optimize-agentic-mcp
Pass
Audited by Gen Agent Trust Hub on May 10, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: Static detectors identified prompt injection patterns within the security reference documentation. These are confirmed false positives, as the strings (e.g., 'ignore previous instructions') are used as educational examples in a guide explaining how to detect and mitigate attacks in MCP server development.
- [SAFE]: The skill uses codebase exploration tools such as 'tree', 'rg', and 'grep' to identify MCP entry points, tool registrations, and schema definitions. These operations are scoped to relevant project directories and do not target sensitive system configuration files or user credentials.
- [SAFE]: The skill recommends several companion tools (e.g., 'build-mcp-use-server', 'test-by-mcpc-cli') for specific tasks like server scaffolding and testing. These tools originate from the author's own repository and are presented as opt-in developer utilities within a consistent vendor namespace.
- [SAFE]: All network operations mentioned in the reference documents, such as fetching specifications from the Official MCP Registry or verifying OAuth metadata, target well-known infrastructure and are documented as standard protocol compliance steps.
Audit Metadata